{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Working with temporal data\n", "\n", "> “Ever since the dawn of civilization, people have not been content to see events as unconnected and inexplicable. They have craved an understanding of the underlying order in the world.\" - Stephen Hawking (1988)\n", "\n", "Time is one of the most fundamental ways how we humans organize things in life and what we use to find understanding of the underlying world, as Stephen {cite}`Hawking1988` famously put it in his book \"A brief history of time\". Hence, it is not surprising that the time dimension is very commonly attached to almost all data that we have in the world (the other dimension is naturally space or location, which we will focus in Part II). Hence, being able to handle and work with temporal information is extremely important when doing data analysis. Time information in the data allows us to see patterns through time (trends) as well as to make predictions into the future (at varying level of confidence). In this section, we will introduce some of the core ideas and functionalities how you can work with temporal data in Python and pandas." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Date and time basics\n", "\n", "Before doing any progamming, we need to understand a few details about time conventions themselves. There are a couple of ways how the time information (commonly referred as `timestamp`) is typically represented. The first one is to represent the temporal information as text with specific format such as `\"2020-12-24T09:12:45Z\"` which follows an international ISO 8601 standard [^isostandard]. In this format, the date components (year, month, day) are separated with dash (`-`) and the time components (hour, minute, second) are separated with colon (`:`). Typically there is a letter `T` separating the date and time components from each other (in some other formats, there could just be a space between them). \n", "\n", "The letter `Z` at the end of the string relates to time zones and means that the time here is represented as Coordinated Universal Time (UTC). Time zone information around the world are expressed using negative or positive offsets from the UTC time standard. `UTC±00:00` is the same as Greenwhich Mean Time (GMT) used e.g. in UK (during winter months), and it was chosen after a series of conferences between 1881-1884 (Ogle, 2015) as the worldwide standard for representing the zone 0 to which all other time zones in the world are relative to. For instance, the local time in Finland is two hours ahead of UTC, meaning that the time zone information is expressed as `UTC+2` or `UTC+02:00`, whereas the local time in New York is four hours behind UTC respectively (i.e. `UTC-4`). \n", "\n", "Another commonly used way to represent time in addition to ISO standardized timestamp is to use a so called *Unix time* (also known as *Epoch time* or *POSIX time*). Unix time is expressed as number of seconds since *Unix Epoch* that was on the first of January 1970 at midnight in UTC (i.e. `1970-01-01T00:00:00Z`). This system was developed to describe a point in time in numerical format, and it is widely used in computers by different operating systems. Unix time increments every second and e.g. the Unix time of December 22nd 2020 at 15:00:00 (UTC) is represented as a single number `1608649200`. Using this kind of numerical representation of time makes it much easier for computers to store and manipulate the time information compared to having the information in textual format (`2020-12-22T15:00:00Z`). \n", "\n", "Lastly, an important convention that relates to time and time zones is a concept called Daylight Saving Time (DST), also known as summer time. DST is a commonly used practice for advancing clocks typically by an hour during warmer months so that darkness falls at a later clock time. This is also the reason why e.g. the time zone information in Finland has an offset of +03:00 (summer time) instead of +02:00 which follows the official time zone specification for Finland. DST became a common practice after the global energy crisis in the 1970s and it was introduced in many countries during the following decade. However, some countries have never followed this convention, and there is an ongoing discussion to abandon the convention e.g. in EU." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Working with temporal data in Python\n", "\n", "Next we will learn how temporal data can be handled in Python in general. The most fundamental Python library for dealing with temporal data is called `datetime` that supplies functionalities to construct and manipulate dates and times. The module is one of the Python's standard libraries, hence you do not need to install it separately. Other useful libraries for dealing with time and dates are `time` and `calendar` which provide some additional functionalities in addition to the functionalities provided by `datetime`. In addition, `pytz` library provides many useful functionalities to deal with time zone information and `dateutil` provides some handy functionalities to automate date parsing. Most often the `datetime` module and `pytz` (in addition to pandas) provide everything you need when doing analysis with temporal data. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Constructing datetime objects\n", "\n", "A `datetime` object is a representation of time in a way that Python can understand and operate with it. With the `datetime` library it is possible to construct a `datetime` object for example by `parsing` it from text following the ISO 8601 format or from the Unix timestamp (notice that the name of the library and the object are the same). Let's see how we can construct a `datetime` object from text using a function `strptime()`. The `strptime()` function takes the variable `timestamp` as the first argument representing the time in textual format and a second argument (\"%Y-%m-%dT%H:%M:%S\") which looks quite complicated, but it is the `format code` that instructs the `datetime` how the textual representation of the time should be interpreted:" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "datetime.datetime(2020, 12, 22, 15, 0)" ] }, "execution_count": 1, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from datetime import datetime\n", "\n", "timestamp = \"2020-12-22T15:00:00\"\n", "datetime.strptime(timestamp, \"%Y-%m-%dT%H:%M:%S\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As a result we get a `datetime` object in which the date has been converted into a set of numbers, in which the first one is the year, second one is month, third one day, and so on. This structure is always the same, no matter in what format the input data has been. By using the format codes (i.e. the second argument of the function), we can tell the computer that the first component in our string represents year in 4 digits. The `%Y` characters (i.e. a *directive*) is a special way to instruct the computer to do so. In a similar manner, we can instruct the computer to read the month based on numbers with two digits (`12`) using the directive `%m`. As you can see, there is always a `%` character and then some letter after it, which is the way how we determine how the date is formatted in our text. Notice that the date components are separated with dash (`-`) and the time components are separated with colons (`:`). As we can see from our format code, we can add those separator characters between the directives just as they are represented in the timestamp. Following this approach it is possible to parse basically any kind of time information represented as text. A few of the most commonly used datetime format code directives are listed in Table 3.1. \n", "\n", ": _**Table 3.1**. Most commonly used datetime directives and their meaning._\n", "\n", "|Directive |Description |Examples |\n", "|:---------|:------------------------------|:----------------------|\n", "|`%y` |Year without century as a |00, 20, etc. |\n", "| |zero-padded number | |\n", "|`%Y` |Year including century as a |2000, 1920, etc. |\n", "| |zero-padded number | |\n", "|`%m` |Month as a zero-padded number |01, ..., 12 |\n", "|`%b` |Month as abbreviated name |Jan, ..., Dec |\n", "|`%B` |Month as full name |January, ..., December |\n", "|`%d` |Day of the month as a |01, ..., 31 |\n", "| |zero-padded number | |\n", "|`%H` |Hour (24-hour clock) as a |00, ..., 23 |\n", "| |zero-padded number | |\n", "|`%I` |Hour (12-hour clock) as a |01, ..., 12 |\n", "| |zero-padded number | |\n", "|`%p` |AM or PM (12-hour clock) |am, pm, AM, PM |\n", "|`%M` |Minute as a zero-padded number |00, ..., 59 |\n", "|`%S` |Second as a zero-padded number |00, ..., 59 |\n", "|`%z` |UTC offset (can be empty) |+0000, +2000, etc. |\n", "|`%Z` |Time zone name (can be empty) |UTC, GMT |\n", "|`%a` |Weekday as abbreviated name |Mon, ..., Sun |\n", "|`%A` |Weekday as full name |Monday, ..., Sunday |\n", "|`%w` |Weekday as a number where 0 is |0, ..., 6 |\n", "| |Sunday and 6 is Saturday | |\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "To crystallize the understanding how the timestamps can be parsed, let's look at a few more examples where we also include time zone information and parse the date information based on a format how we typically write dates as humans (without time component). Let's start by adding `+0200` to the end of the timestamp to express the UTC+2 time zone: " ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "datetime.datetime(2020, 12, 22, 15, 0, tzinfo=datetime.timezone(datetime.timedelta(seconds=7200)))" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "timestamp_with_tz = \"2020-12-22T15:00:00 +0200\"\n", "dtz = datetime.strptime(timestamp_with_tz, \"%Y-%m-%dT%H:%M:%S %z\")\n", "dtz" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we can see, now we produced the `datetime` object having time zone information attached into the `tzinfo` attribute showing the offset (i.e. *timedelta*) from UTC represented in seconds. Having the timezone information attached can be very useful if doing analysis with temporal data that has been collected from different parts of the world (under different time zones). Let's still take a look at an example in which we parse the `datetime` object from a textual representation that is written in a way how we humans normally write dates:" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "datetime.datetime(2020, 12, 22, 0, 0)" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "date_written = \"22 December 2020\"\n", "datetime.strptime(date_written, \"%d %B %Y\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Here, we used a bit different format for instructing the computer how to read the text by using `%B` to denote a written month name, also we changed the order of the directives and used an empty space between them. Writing the exact form of how the timestamp information is written can be quite devious work, especially if working with handwritten data where stucture of the time information can vary a lot. Luckily, Python provides a handy library `dateutil` that automates most of this work. We can use a function called `parse()` that can automatically construct the `datetime` object from various formats:" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2020-12-22 15:00:00+02:00 \n", " 2020-12-22 00:00:00\n" ] } ], "source": [ "from dateutil.parser import parse\n", "\n", "timestamp_with_tz = \"2020-12-22T15:00:00 +0200\"\n", "date_written = \"22 December 2020\"\n", "dt1 = parse(timestamp_with_tz)\n", "dt2 = parse(date_written)\n", "print(dt1, \"\\n\", dt2)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "All of the previous examples focused on using textual representation of time as input. Naturally it is also possible to construct a datetime object using Unix time representation. Because Unix time is a simple number, it is much more straightforward to parse a datetime object based on them. This is how would convert a Unix time `1608649200` into a `datetime` using `fromtimestamp()` function:" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "datetime.datetime(2020, 12, 22, 17, 0)" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "unix_time = 1608649200\n", "datetime.fromtimestamp(unix_time)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Wait, what? In the beginning of this chapter, we said that Unix time `1608649200` should be \"December 22nd 2020 at 15:00:00 (UTC)\", but here we have the result as five o'clock (i.e. 17, 0). This is due to the fact that the computer that this code is written with is located at a Finnish time zone (UTC+02), and as the Unix time is always passed as UTC, the computer automatically returns the timestamp in local timezone (i.e. in Finnish time or whatever the local time has been specified in your computer's settings). This is a good thing to understand because it can be rather confusing sometimes when working with temporal data. Finally, it is naturally possible to initialize the `datetime` object also directly by passing integer numbers into the object itself. Here we use the same date and time as before but add `30` seconds to the end: " ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "datetime.datetime(2020, 12, 22, 17, 0, 30)" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "dt = datetime(2020, 12, 22, 17, 0, 30)\n", "dt" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Parsing temporal components from datetime object\n", "\n", "There are few useful tricks that you can do with the `datetime` objects. It is for example possible to parse specific temporal components directly from the `datetime` object. You can for example access the `.year`, `.month`, `.day`, `.hour`, `.minute` and `.second` attributes very easily that will return the value of the temporal element as an integer number:" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2020\n", "12\n", "22\n", "17\n" ] } ], "source": [ "dt = datetime(2020, 12, 22, 17, 0, 30)\n", "print(dt.year)\n", "print(dt.month)\n", "print(dt.day)\n", "print(dt.hour)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In addition, you can parse for example the day of week by using `weekday()` function or parse only the date components from the `datetime` object by using `date()` function:" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "1\n", "2020-12-22\n" ] } ], "source": [ "print(dt.weekday())\n", "print(dt.date())" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "By using `strftime()` function, you can use the same directives from Table 2.1 to construct and convert the `datetime` object to any text representation of the date and time. Next, we will see how this function can be used quite nicely for creating an easily understandable text based on our date and time information:" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "'22 December 2020 is Tuesday. The time is 05:00 PM.'" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "custom_format = \"%d %B %Y is %A. The time is %I:%M %p.\"\n", "datetime.strftime(dt, custom_format)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Working with temporal data in pandas\n", "\n", "After learning the basics of how the temporal data can be represented as `datetime` objects, we can continue and see how temporal data can be manipulated in pandas. Pandas is extremely powerful and flexible what comes to working with temporal data. For instance, all the simple tricks that we saw earlier with `datetime` can be done also with pandas. When doing data analysis with pandas, it is typical that instead of dealing with single time observations, you actually work with time series data, such as the hourly temperature values that we analyzed in the previous section. Pandas supports many useful functionalities related to parsing, manipulating and aggregating time series data. Let's start exploring the temporal capabilities of pandas by reading the hourly temperature observations from the same CSV file in a similar manner as we did previously. As a reminder, this is how the data look like:\n", "\n", "``` \n", " USAF WBAN YR--MODAHRMN DIR SPD GUS CLG SKC L M H VSB MW MW MW MW AW ...\n", "029440 99999 190601010600 090 7 *** *** OVC * * * 0.0 ** ** ** ** ** ...\n", "029440 99999 190601011300 *** 0 *** *** OVC * * * 0.0 ** ** ** ** ** ...\n", "029440 99999 190601012000 *** 0 *** *** OVC * * * 0.0 ** ** ** ** ** ...\n", "029440 99999 190601020600 *** 0 *** *** CLR * * * 0.0 ** ** ** ** ** ...\n", "```\n", "\n", "The timestamps stored in the column `YR--MODAHRMN` can be automatically converted to `datetime` objects when reading the data by using `parse_dates` parameter and providing a list of column names that should be parsed to `datetime`. We also specify with `usecols` parameter that we only want to keep three columns from the file (station number, time information and temperature):" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTIMETEMP_F
0298201906-01-01 06:00:0034.0
1298201906-01-01 13:00:0032.0
2298201906-01-01 20:00:0030.0
3298201906-01-02 06:00:0033.0
4298201906-01-02 13:00:0035.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TIME TEMP_F\n", "0 29820 1906-01-01 06:00:00 34.0\n", "1 29820 1906-01-01 13:00:00 32.0\n", "2 29820 1906-01-01 20:00:00 30.0\n", "3 29820 1906-01-02 06:00:00 33.0\n", "4 29820 1906-01-02 13:00:00 35.0" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import pandas as pd\n", "\n", "# Read the data and parse dates\n", "fp = \"data/029820.txt\"\n", "data = pd.read_csv(\n", " fp,\n", " delim_whitespace=True,\n", " na_values=[\"*\", \"**\", \"***\", \"****\", \"*****\", \"******\"],\n", " usecols=[\"USAF\", \"YR--MODAHRMN\", \"TEMP\"],\n", " parse_dates=[\"YR--MODAHRMN\"],\n", ")\n", "# Rename the columns\n", "new_cols = [\"STATION_NUMBER\", \"TIME\", \"TEMP_F\"]\n", "data.columns = new_cols\n", "data.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we can see, the values in `TIME` column were automatically parsed into a format that are much easier to read than the original ones. When using pandas datetime parsing functionalities, the timestamps can be represented in different formats such as text in ISO 8601 format or Unix times (seconds since 1.1.1970). Let's take a look how the values in our `TIME` column look like:" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0 1906-01-01 06:00:00\n", "1 1906-01-01 13:00:00\n", "2 1906-01-01 20:00:00\n", "3 1906-01-02 06:00:00\n", "4 1906-01-02 13:00:00\n", "Name: TIME, dtype: datetime64[ns]" ] }, "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data[\"TIME\"].head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we can see, pandas actually converted the timestamp values into a time series having specific data type called `datetime64`. The individual value of the time series is a specific pandas object called `Timestamp` that is a similar object as the Python's regular `datetime` object that we saw previously: " ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Timestamp('1906-01-01 06:00:00')" ] }, "execution_count": 12, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data.loc[0, \"TIME\"]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The `TimeStamp` object contains all the same attributes as the `datetime` object, but in addition, it has many useful attributes and functions that can be used in a vectorized manner. For instance, you can parse different temporal components from the time series stored in a column very efficiently. Let's parse the temporal components from `TIME` into columns `YEAR`, `MONTH`, `DAY` and `HOUR`: " ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTIMETEMP_FYEARMONTHDAYHOUR
0298201906-01-01 06:00:0034.01906116
1298201906-01-01 13:00:0032.019061113
2298201906-01-01 20:00:0030.019061120
3298201906-01-02 06:00:0033.01906126
4298201906-01-02 13:00:0035.019061213
\n", "
" ], "text/plain": [ " STATION_NUMBER TIME TEMP_F YEAR MONTH DAY HOUR\n", "0 29820 1906-01-01 06:00:00 34.0 1906 1 1 6\n", "1 29820 1906-01-01 13:00:00 32.0 1906 1 1 13\n", "2 29820 1906-01-01 20:00:00 30.0 1906 1 1 20\n", "3 29820 1906-01-02 06:00:00 33.0 1906 1 2 6\n", "4 29820 1906-01-02 13:00:00 35.0 1906 1 2 13" ] }, "execution_count": 13, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data[\"YEAR\"] = data[\"TIME\"].dt.year\n", "data[\"MONTH\"] = data[\"TIME\"].dt.month\n", "data[\"DAY\"] = data[\"TIME\"].dt.day\n", "data[\"HOUR\"] = data[\"TIME\"].dt.hour\n", "data.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now the new columns have corresponding integer values for each temporal component. We were able to access all the temporal attributes by using the `.dt` accessor that is available for all `Series` having `datetime64` data type. We can also combine the datetime functionalities with other methods from pandas. For example, we can check that we have observations from 69 years in total: " ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "69" ] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data[\"TIME\"].dt.year.nunique()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Handling time zones in pandas" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "When reading time series data to pandas as we did previously, the timestamp objects are represented without time zone information. However, it is straighforward to attach time zone information to a Series with Timestamp objects by using a method called `tz_localize()`. When manipulating the data based on time information in pandas, the first step that you typically do, is to set a *{term}`DatetimeIndex`* for you DataFrame (here based on column 'TIME') with `set_index()` method. Next, we will set the index and specify appropriate time zone for our data (*Europe/Helsinki*). We will also specify the `nonexistent` parameter as \"shift_forward\", due to the fact that when the clocks are adjusted when changing from winter time to summer time, there is literally one non-existing hour during the night. With this parameter we tell pandas that this is okay, and we just want to shift the time forward (by an hour in most cases) in case there are such \"non-existing\" time observations. With `ambiguous` parameter, we tell pandas that when adjusting the clocks backwards (moving from summer time to winter time), the observations having ambigous times (e.g. duplicate timestamps due to Daylight Saving Time) are set as `NaT` (Not-a-Time). Notice that it is up to you to decide what is appropriate way of handling such cases (there are also other options for this parameter, such as \"infer\"): " ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_FYEARMONTHDAYHOUR
TIME
2019-10-01 19:00:00+03:002982047.0201910119
2019-10-01 20:00:00+03:002982047.0201910120
2019-10-01 21:00:00+03:002982047.0201910121
2019-10-01 22:00:00+03:002982047.0201910122
2019-10-01 23:00:00+03:002982046.0201910123
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F YEAR MONTH DAY HOUR\n", "TIME \n", "2019-10-01 19:00:00+03:00 29820 47.0 2019 10 1 19\n", "2019-10-01 20:00:00+03:00 29820 47.0 2019 10 1 20\n", "2019-10-01 21:00:00+03:00 29820 47.0 2019 10 1 21\n", "2019-10-01 22:00:00+03:00 29820 47.0 2019 10 1 22\n", "2019-10-01 23:00:00+03:00 29820 46.0 2019 10 1 23" ] }, "execution_count": 15, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data = data.set_index(\"TIME\")\n", "data = data.tz_localize(\"Europe/Helsinki\", nonexistent=\"shift_forward\", ambiguous=\"NaT\")\n", "data.tail()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now the TIME column is set as our index and it was moved to the left and replaced the original sequantial numbers. Notice that the `TIME` column is no longer part of the columns in our DataFrame (we will later see how to get it back as a column). Also the time zone information was attached to the Timestamp objects as can be seen by the five last characters showing `+03:00`. From this information we can see that the time zone in Finland is +3 hours at the given date (summer time). We can easily access the values in the index by calling it as follows: " ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "DatetimeIndex(['1906-01-01 06:00:00+01:40', '1906-01-01 13:00:00+01:40',\n", " '1906-01-01 20:00:00+01:40', '1906-01-02 06:00:00+01:40',\n", " '1906-01-02 13:00:00+01:40', '1906-01-02 20:00:00+01:40',\n", " '1906-01-03 06:00:00+01:40', '1906-01-03 13:00:00+01:40',\n", " '1906-01-03 20:00:00+01:40', '1906-01-04 06:00:00+01:40',\n", " ...\n", " '2019-10-01 14:00:00+03:00', '2019-10-01 15:00:00+03:00',\n", " '2019-10-01 16:00:00+03:00', '2019-10-01 17:00:00+03:00',\n", " '2019-10-01 18:00:00+03:00', '2019-10-01 19:00:00+03:00',\n", " '2019-10-01 20:00:00+03:00', '2019-10-01 21:00:00+03:00',\n", " '2019-10-01 22:00:00+03:00', '2019-10-01 23:00:00+03:00'],\n", " dtype='datetime64[ns, Europe/Helsinki]', name='TIME', length=198334, freq=None)" ] }, "execution_count": 16, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data.index" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "Timestamp('1906-01-01 06:00:00+0140', tz='Europe/Helsinki')" ] }, "execution_count": 17, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data.index[0]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we can see, the index is now a `DatetimeIndex` which is essentially an array consisting of `Timestamp` objects armed with all the attributes and methods that we have seen in the previous examples. In addition, our array has time zone specified as \"Europe/Helsinki\". Hint: If you want to know the available time zones, you can access them by `from pytz import all_timezones` which imports them into the variable `all_timezones` as a list. \n", "\n", "There are a couple of important notions that we can make from the data in this array. Firstly, the Timestamps in the index from the year 1906 seem to have a quite odd time zone information attached to them (+01:40), which differs from the offset of the last observations (+03:00) by an hour and twenty minutes. This might seem a bit strange, although everything is as it should, and a couple of international conventions regarding time explain these differences. First explanation relates to a fact that representing local time following time zone specification as an offset from UTC is merely a contract between countries targeted to standardize the representation of time in different parts of the world. But as with any international contract, the standardization process typically happens at different times in different countries, and it is not unheard of to make changes to these agreements. An example of such a change can be witnessed here, as Finland changed its time zone from Helsinki Mean Time (with +01:40 offset from UTC) to Eastern European Time (with +02:00 from UTC) in May 1st 1921. Hence, all timestamps recorded before the aforementioned date have a somewhat unusual offset from UTC (100 minutes). Similar special arrangements have been quite common in the history, so it is good to be careful and aware of these situations whenever working with historical time series. Second important convention relates to the Daylight Saving Time (DST), which was introduced in Finland in 1981, and is the reason why the last values in our DatetimeIndex have an offset of +03:00 (summer time) instead of +02:00. As you might guess, all these international practices and contracts can make working with date and time information quite tricky. Luckily, pandas (and the underlying pytz library) handles these things very well, and we do not need to worry too much about the international time convensions. \n", "\n", "In some occasions, you might want to convert your time series data from one time zone to another. If you for example, would like to understand what the time was in New York when these weather observations were recorded, you could convert the time information easily with `.dt.tz_convert()` function. Before making the conversion between the time zones, we need to return the TIME Series from the index back to a normal pandas column by using `.reset_index()` method that returns the `TIME` column back to DataFrame as a normal column. When manipulating time series, it is quite typical to set a specific column as an index for the DataFrame and then swap it back to a \"normal\" column for further manipulation, as in the following: " ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TIMESTATION_NUMBERTEMP_FYEARMONTHDAYHOURNY_TIME
01906-01-01 06:00:00+01:402982034.019061161905-12-31 23:20:00-05:00
11906-01-01 13:00:00+01:402982032.0190611131906-01-01 06:20:00-05:00
21906-01-01 20:00:00+01:402982030.0190611201906-01-01 13:20:00-05:00
31906-01-02 06:00:00+01:402982033.019061261906-01-01 23:20:00-05:00
41906-01-02 13:00:00+01:402982035.0190612131906-01-02 06:20:00-05:00
\n", "
" ], "text/plain": [ " TIME STATION_NUMBER TEMP_F YEAR MONTH DAY HOUR \\\n", "0 1906-01-01 06:00:00+01:40 29820 34.0 1906 1 1 6 \n", "1 1906-01-01 13:00:00+01:40 29820 32.0 1906 1 1 13 \n", "2 1906-01-01 20:00:00+01:40 29820 30.0 1906 1 1 20 \n", "3 1906-01-02 06:00:00+01:40 29820 33.0 1906 1 2 6 \n", "4 1906-01-02 13:00:00+01:40 29820 35.0 1906 1 2 13 \n", "\n", " NY_TIME \n", "0 1905-12-31 23:20:00-05:00 \n", "1 1906-01-01 06:20:00-05:00 \n", "2 1906-01-01 13:20:00-05:00 \n", "3 1906-01-01 23:20:00-05:00 \n", "4 1906-01-02 06:20:00-05:00 " ] }, "execution_count": 18, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 'TIME' column is returned back as a normal column\n", "data = data.reset_index()\n", "data[\"NY_TIME\"] = data[\"TIME\"].dt.tz_convert(\"US/Eastern\")\n", "data.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now we can see, that the timestamps were converted to represent the times in US/Eastern time zone and stored to the NY_TIME column. For example the first observation in our data that was recorded at 6 AM 1st of January in 1906 was correctly converted to a value from the last day of previous year at 11:20 PM (`1905-12-31 23:20`). This functionality can be very useful when working with temporal data from different parts of the world. Quite often the data collected from different global services (such as tweets collected from Twitter) store the information as UTC values. Hence, it is up to the user to parse the correct local time for tweets posted in different parts of the world. Using pandas for doing these kind of manipulation with the temporal data is extremely handy and efficient. In case your data is already represented in different timezones, pandas makes your life easy because timestamps are stored under the hood always in UTC. This means that it is straightforward to do operations between the two time series without any additional timezone conversions. We can for example do simple calculations between two DataFrames that are represented in different timezones: " ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TEMP_F
TIME
1906-01-01 06:00:00+01:4034.0
1906-01-01 13:00:00+01:4032.0
1906-01-01 20:00:00+01:4030.0
1906-01-02 06:00:00+01:4033.0
1906-01-02 13:00:00+01:4035.0
\n", "
" ], "text/plain": [ " TEMP_F\n", "TIME \n", "1906-01-01 06:00:00+01:40 34.0\n", "1906-01-01 13:00:00+01:40 32.0\n", "1906-01-01 20:00:00+01:40 30.0\n", "1906-01-02 06:00:00+01:40 33.0\n", "1906-01-02 13:00:00+01:40 35.0" ] }, "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample1 = data[[\"TIME\", \"TEMP_F\"]].head()\n", "sample1 = sample1.set_index(\"TIME\")\n", "sample1" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TEMP_F
NY_TIME
1905-12-31 23:20:00-05:0034.0
1906-01-01 06:20:00-05:0032.0
1906-01-01 13:20:00-05:0030.0
1906-01-01 23:20:00-05:0033.0
1906-01-02 06:20:00-05:0035.0
\n", "
" ], "text/plain": [ " TEMP_F\n", "NY_TIME \n", "1905-12-31 23:20:00-05:00 34.0\n", "1906-01-01 06:20:00-05:00 32.0\n", "1906-01-01 13:20:00-05:00 30.0\n", "1906-01-01 23:20:00-05:00 33.0\n", "1906-01-02 06:20:00-05:00 35.0" ] }, "execution_count": 20, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample2 = data[[\"NY_TIME\", \"TEMP_F\"]].head()\n", "sample2 = sample2.set_index(\"NY_TIME\")\n", "sample2" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now we have the same temperature values in both DataFrames but these observations are represented in different time zones. Now, we can for instance sum up these temperatures, which works because the time of the observations matches:" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "TIME\n", "1906-01-01 04:20:00+00:00 68.0\n", "1906-01-01 11:20:00+00:00 64.0\n", "1906-01-01 18:20:00+00:00 60.0\n", "1906-01-02 04:20:00+00:00 66.0\n", "1906-01-02 11:20:00+00:00 70.0\n", "Name: TEMP_F, dtype: float64" ] }, "execution_count": 21, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample1[\"TEMP_F\"] + sample2[\"TEMP_F\"]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As you can see, pandas automatically converted the temporal information to UTC timezone, and the temperature values were summed up correctly." ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Selecting data based on DatetimeIndex" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In chapter 3.3, we saw how to select observations using simple string manipulation operations applied to the timestamp. Although this worked quite well for our specific case, it is a rather glumpsy and inflexible approach to work with temporal information. A much better approach, is to take advantage of the `datetime` objects stored in the time series. Let's again first set the `TIME` column as index for our DataFrame by using the method `set_index()`:" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_F
TIME
1906-01-01 06:00:002982034.0
1906-01-01 13:00:002982032.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F\n", "TIME \n", "1906-01-01 06:00:00 29820 34.0\n", "1906-01-01 13:00:00 29820 32.0" ] }, "execution_count": 22, "metadata": {}, "output_type": "execute_result" } ], "source": [ "fp = \"data/029820.txt\"\n", "data = pd.read_csv(\n", " fp,\n", " delim_whitespace=True,\n", " na_values=[\"*\", \"**\", \"***\", \"****\", \"*****\", \"******\"],\n", " usecols=[\"USAF\", \"YR--MODAHRMN\", \"TEMP\"],\n", " parse_dates=[\"YR--MODAHRMN\"],\n", ")\n", "# Rename the columns\n", "new_cols = [\"STATION_NUMBER\", \"TIME\", \"TEMP_F\"]\n", "data.columns = new_cols\n", "\n", "# Set TIME as index\n", "data = data.set_index(\"TIME\")\n", "data.head(2)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now as we have equipped our DataFrame with *DatetimeIndex*, it is easy to filter the data based on specific criteria and e.g. select all observations that have been recorded in specific months of 2019. To do this, we can use `.loc` indexing in a similar manner as before, but pass the criteria for selecting the data using either `datetime` objects or strings that can be interpreted as a date. For instance, we can select the observations for January 2019 in a following manner based on datetime labels: " ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "First obs: 2019-01-01 00:00:00 \n", "Last obs: 2019-01-31 00:00:00\n" ] } ], "source": [ "start_dt = datetime(2019, 1, 1)\n", "end_dt = datetime(2019, 1, 31)\n", "january = data.loc[start_dt:end_dt].copy()\n", "first, last = january.index.min(), january.index.max()\n", "print(\"First obs:\", first, \"\\nLast obs:\", last)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we mentioned, pandas also provides a nice and easy way to select observations e.g. for specific month by specifying the selection criteria as a string. This approach is very convenient because we can select observations by passing a single label to the `loc` indexer and pandas automatically detects based on the string, what should be done: " ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "First obs: 2019-02-01 00:00:00 \n", "Last obs: 2019-02-28 23:00:00\n" ] } ], "source": [ "february = data.loc[\"2019-02\"].copy()\n", "first, last = february.index.min(), february.index.max()\n", "print(\"First obs:\", first, \"\\nLast obs:\", last)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As a result, we got all observations that were recorded during February 2019. Using this approach, called *partial string indexing*, is useful because it makes it easy for us to deal with irregular number of days in February (leap year). By passing the time window as a string that consists of the year and month information, we do not need to worry about leap years ourselves, as pandas does this automatically for us:" ] }, { "cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "First obs: 2016-02-01 00:00:00 \n", "Last obs: 2016-02-29 23:00:00\n" ] } ], "source": [ "leap_year = data.loc[\"2016-02\"].copy()\n", "first, last = leap_year.index.min(), leap_year.index.max()\n", "print(\"First obs:\", first, \"\\nLast obs:\", last)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Following the same approach, we can adjust the criteria to be a more specific time range, such as first two weeks of February:" ] }, { "cell_type": "code", "execution_count": 26, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_F
TIME
2018-02-14 22:00:002982033.0
2018-02-14 23:00:002982033.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F\n", "TIME \n", "2018-02-14 22:00:00 29820 33.0\n", "2018-02-14 23:00:00 29820 33.0" ] }, "execution_count": 26, "metadata": {}, "output_type": "execute_result" } ], "source": [ "first_two_weeks = data.loc[\"2018-02-01\":\"2018-02-14\"].copy()\n", "first_two_weeks.tail(2)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Or select data covering the whole year of 2018: " ] }, { "cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_F
TIME
2018-01-01 00:00:002982034.0
2018-01-01 01:00:002982033.0
2018-01-01 02:00:002982034.0
2018-01-01 03:00:002982034.0
2018-01-01 04:00:002982035.0
.........
2018-12-31 19:00:002982036.0
2018-12-31 20:00:002982036.0
2018-12-31 21:00:002982037.0
2018-12-31 22:00:002982038.0
2018-12-31 23:00:002982039.0
\n", "

8439 rows × 2 columns

\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F\n", "TIME \n", "2018-01-01 00:00:00 29820 34.0\n", "2018-01-01 01:00:00 29820 33.0\n", "2018-01-01 02:00:00 29820 34.0\n", "2018-01-01 03:00:00 29820 34.0\n", "2018-01-01 04:00:00 29820 35.0\n", "... ... ...\n", "2018-12-31 19:00:00 29820 36.0\n", "2018-12-31 20:00:00 29820 36.0\n", "2018-12-31 21:00:00 29820 37.0\n", "2018-12-31 22:00:00 29820 38.0\n", "2018-12-31 23:00:00 29820 39.0\n", "\n", "[8439 rows x 2 columns]" ] }, "execution_count": 27, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data.loc[\"2018\"]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Combining this kind of indexing with methods such as `first()` or `last()` is extremely powerful, as you can easily retrieve for example all observations within first 6 hours and 10 minutes of a specific day (an arbitrary duration), in case you would be interested to understand temperatures during night time:" ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_F
TIME
2018-01-01 00:00:002982034.0
2018-01-01 01:00:002982033.0
2018-01-01 02:00:002982034.0
2018-01-01 03:00:002982034.0
2018-01-01 04:00:002982035.0
2018-01-01 05:00:002982036.0
2018-01-01 06:00:002982036.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F\n", "TIME \n", "2018-01-01 00:00:00 29820 34.0\n", "2018-01-01 01:00:00 29820 33.0\n", "2018-01-01 02:00:00 29820 34.0\n", "2018-01-01 03:00:00 29820 34.0\n", "2018-01-01 04:00:00 29820 35.0\n", "2018-01-01 05:00:00 29820 36.0\n", "2018-01-01 06:00:00 29820 36.0" ] }, "execution_count": 28, "metadata": {}, "output_type": "execute_result" } ], "source": [ "data.loc[\"2018-01-01\"].first(\"6H 10T\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "The `6H` stands for a duration of six hours (number of hours plus letter `H`) and the `10T` is a duration of ten minutes (number of minutes plus `T`), and you can combine those into a single duration as shown above. In a similar manner, you can specify many other *{term}`DateOffsets`*, such as day (`D`) or second (`S`). The full list of all possible keywords (aliases) can be found from pandas documentation [^dateoffsets]. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Shifting - or leading and lagging data" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "One typical task when working with time series data, is to compare two consecutive values to each other. For instance, you might be interested in understanding how much warmer or colder the temperature was between different hours, or compute the percentual difference of how the value of a given stock has changed (to give a couple of examples). For these kind of operations, pandas provides a functionality called \"shifting\" which can be used to move (shift) data backward and forward. Next, we will see this method in action, by calculating the difference between consecutive temperature values. Let's start by taking a small sample covering the data for 2019:" ] }, { "cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_F
TIME
2019-01-01 00:00:002982039.0
2019-01-01 01:00:002982040.0
2019-01-01 02:00:002982040.0
2019-01-01 03:00:002982040.0
2019-01-01 04:00:002982040.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F\n", "TIME \n", "2019-01-01 00:00:00 29820 39.0\n", "2019-01-01 01:00:00 29820 40.0\n", "2019-01-01 02:00:00 29820 40.0\n", "2019-01-01 03:00:00 29820 40.0\n", "2019-01-01 04:00:00 29820 40.0" ] }, "execution_count": 29, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample = data.loc[\"2019\"].copy()\n", "sample.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we can see based on the DatetimeIndex (TIME), we have temperature observations on an hourly basis in this sample. Now, if we want to calculate what is the difference in temperature between consecutive observations, we can take advantage of the method `.shift()` which takes an integer value as an input. If you pass a positive number, the values in a given Series is shifted forward, and if you pass a negative number the values are shifted backwards: " ] }, { "cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "TIME\n", "2019-01-01 00:00:00 NaN\n", "2019-01-01 01:00:00 39.0\n", "2019-01-01 02:00:00 40.0\n", "2019-01-01 03:00:00 40.0\n", "2019-01-01 04:00:00 40.0\n", " ... \n", "2019-10-01 19:00:00 50.0\n", "2019-10-01 20:00:00 47.0\n", "2019-10-01 21:00:00 47.0\n", "2019-10-01 22:00:00 47.0\n", "2019-10-01 23:00:00 47.0\n", "Name: TEMP_F, Length: 6455, dtype: float64" ] }, "execution_count": 30, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Shifting forward\n", "sample[\"TEMP_F\"].shift(1)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Here, we shifted the values forward by one observation, which can be seen e.g. by looking at how the position of value 39.0 was moved by an hour to 01:00. Furthermore, by looking at the numeric values, you can see that the first value in the Series was replaced with NaN. This is because the temperature values were shifted forward by one observation, meaning that there is not any known temperature that could be placed as the first observation. Hence, pandas fills it with NaN. In a similar manner, when the observations are shifted backwards (with argument -1), there is NaN inserted into the Series, but in this case, at the very end of it: " ] }, { "cell_type": "code", "execution_count": 31, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "TIME\n", "2019-01-01 00:00:00 40.0\n", "2019-01-01 01:00:00 40.0\n", "2019-01-01 02:00:00 40.0\n", "2019-01-01 03:00:00 40.0\n", "2019-01-01 04:00:00 40.0\n", " ... \n", "2019-10-01 19:00:00 47.0\n", "2019-10-01 20:00:00 47.0\n", "2019-10-01 21:00:00 47.0\n", "2019-10-01 22:00:00 46.0\n", "2019-10-01 23:00:00 NaN\n", "Name: TEMP_F, Length: 6455, dtype: float64" ] }, "execution_count": 31, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Shifting backward\n", "sample[\"TEMP_F\"].shift(-1)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Another important aspect to notice from these examples, is that in both cases, one observation was removed from the Series. In the first example, the last known temperature (46.0) was dropped from the end of the Series, and in the second example, the first temperature (39.0) was dropped. This is expected behavior, because pandas wants to ensure that the Series keeps its shape, i.e. the length of the Series does not change. This behavior is beneficial, because in this way, it is straightforward to do calculations between the two Series, or e.g. add the shifted values as a column to the same DataFrame:" ] }, { "cell_type": "code", "execution_count": 32, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_Fnext_hour
TIME
2019-01-01 00:00:002982039.040.0
2019-01-01 01:00:002982040.040.0
2019-01-01 02:00:002982040.040.0
2019-01-01 03:00:002982040.040.0
2019-01-01 04:00:002982040.040.0
2019-01-01 05:00:002982040.040.0
2019-01-01 06:00:002982040.040.0
2019-01-01 07:00:002982040.041.0
2019-01-01 08:00:002982041.040.0
2019-01-01 09:00:002982040.040.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F next_hour\n", "TIME \n", "2019-01-01 00:00:00 29820 39.0 40.0\n", "2019-01-01 01:00:00 29820 40.0 40.0\n", "2019-01-01 02:00:00 29820 40.0 40.0\n", "2019-01-01 03:00:00 29820 40.0 40.0\n", "2019-01-01 04:00:00 29820 40.0 40.0\n", "2019-01-01 05:00:00 29820 40.0 40.0\n", "2019-01-01 06:00:00 29820 40.0 40.0\n", "2019-01-01 07:00:00 29820 40.0 41.0\n", "2019-01-01 08:00:00 29820 41.0 40.0\n", "2019-01-01 09:00:00 29820 40.0 40.0" ] }, "execution_count": 32, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample[\"next_hour\"] = sample[\"TEMP_F\"].shift(-1)\n", "sample.head(10)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now the first value in the column `next_hour` actually represents the temperature from 01:00, the second value from 02:00, etc. because we shifted the values backwards. At this point, our data is structured in such a manner that it is very easy to do calculations between these two columns and e.g. calculate the difference between consecutive hours:" ] }, { "cell_type": "code", "execution_count": 33, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_Fnext_hourdiff
TIME
2019-01-01 00:00:002982039.040.01.0
2019-01-01 01:00:002982040.040.00.0
2019-01-01 02:00:002982040.040.00.0
2019-01-01 03:00:002982040.040.00.0
2019-01-01 04:00:002982040.040.00.0
...............
2019-10-01 19:00:002982047.047.00.0
2019-10-01 20:00:002982047.047.00.0
2019-10-01 21:00:002982047.047.00.0
2019-10-01 22:00:002982047.046.0-1.0
2019-10-01 23:00:002982046.0NaNNaN
\n", "

6455 rows × 4 columns

\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F next_hour diff\n", "TIME \n", "2019-01-01 00:00:00 29820 39.0 40.0 1.0\n", "2019-01-01 01:00:00 29820 40.0 40.0 0.0\n", "2019-01-01 02:00:00 29820 40.0 40.0 0.0\n", "2019-01-01 03:00:00 29820 40.0 40.0 0.0\n", "2019-01-01 04:00:00 29820 40.0 40.0 0.0\n", "... ... ... ... ...\n", "2019-10-01 19:00:00 29820 47.0 47.0 0.0\n", "2019-10-01 20:00:00 29820 47.0 47.0 0.0\n", "2019-10-01 21:00:00 29820 47.0 47.0 0.0\n", "2019-10-01 22:00:00 29820 47.0 46.0 -1.0\n", "2019-10-01 23:00:00 29820 46.0 NaN NaN\n", "\n", "[6455 rows x 4 columns]" ] }, "execution_count": 33, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample[\"diff\"] = sample[\"next_hour\"] - sample[\"TEMP_F\"]\n", "sample" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As you can see, the last value in column `diff` is NaN, which expected as we don't have information about the following hour on the last row. \n", "\n", "In the previous examples, we did *naive* shifts forward and backward by merely changing the position of the observations. Naturally, it is also possible to take advantage of hour DatetimeIndex, and do more sophisticated shifts based on time. We can for example compare the temperatures between consecutive days or weeks, for example. Let's see how we can calculate the weekly differences in the temperatures using the same `.shift()` method by adding a parameter `freq` which accepts a *DateOffset* as an argument (that we already saw and used in the previous section). This time, we calculate the difference directly without storing the temperatures in a separate column:" ] }, { "cell_type": "code", "execution_count": 34, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_Fnext_hourdiffweekly_diff
TIME
2019-01-01 00:00:002982039.040.01.0-2.0
2019-01-01 01:00:002982040.040.00.0-4.0
2019-01-01 02:00:002982040.040.00.0-3.0
2019-01-01 03:00:002982040.040.00.0-3.0
2019-01-01 04:00:002982040.040.00.0-3.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F next_hour diff weekly_diff\n", "TIME \n", "2019-01-01 00:00:00 29820 39.0 40.0 1.0 -2.0\n", "2019-01-01 01:00:00 29820 40.0 40.0 0.0 -4.0\n", "2019-01-01 02:00:00 29820 40.0 40.0 0.0 -3.0\n", "2019-01-01 03:00:00 29820 40.0 40.0 0.0 -3.0\n", "2019-01-01 04:00:00 29820 40.0 40.0 0.0 -3.0" ] }, "execution_count": 34, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample[\"weekly_diff\"] = sample[\"TEMP_F\"].shift(-7, freq=\"D\") - sample[\"TEMP_F\"]\n", "sample.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Here, we specified -7 as the first argument and passed a letter \"D\" for the `freq` parameter, meaning that we want shift the observations by 7 days. Then, we calculated the difference between the temperature on the following week and stored the result in a separate column `\"weekly_diff\"`. In a similar manner, you can easily calculate differences between observations taken at different times, which can be very efficient and useful to understand changes in time series. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Resampling and calculating rolling statistics" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Now as we know how to take advantage of the DatetimeIndex to make selections and time zone conversions as well as manipulate data by shifting observations, we are ready to explore further the functionalities of pandas for working with temporal data. One typical task when doing analysis with time series data is to *{term}`resample`* the time series from one frequency to another, such as aggregating the hourly temperature observations to daily averages. This might sound familiar, as we did this process previously (in 3.3) by using specific string manipulation tricks, and then grouping and aggregating the values. Now we will learn how to use pandas resampling functionalities that allows to conduct both downsampling (i.e. aggregating values) and upsampling based on specific interpolation methods. When doing interpolation, you fill missing values between observations based on specific criteria, such as taking the average between two consecutive observations. For instance, you could interpolate hourly temperature observations into 30 minute frequencies by assuming that the temperature in the middle of two time steps, such as 14:00 and 15:00, is the average of both giving you an estimate what the temperature might be at 14:30. Notice that whenever interpolating/upsampling data, you are basically making \"sophisticated guesses\" and there is always uncertainty involved with your estimates. \n", "\n", "Let's start by downsampling our observations into daily averages. To do this, we can use a `resample()` method that works in a bit similar manner as `groupby()` but it works with `DatetimeIndex` allowing you to easily control how the values should be grouped. Notice, that this only works with time series data, i.e. you need to make sure that you have set `DatetimeIndex` for your DataFrame. Whenever resampling data, you typically want to combine it with a specific aggregation method, such as `mean()`, `max()`, `min()`, which will then specify how you want to aggregate the data, such as calculating the daily averages: " ] }, { "cell_type": "code", "execution_count": 35, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_F
TIME
1906-01-0129820.032.000000
1906-01-0229820.032.000000
1906-01-0329820.027.666667
1906-01-0429820.031.333333
1906-01-0529820.031.666667
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F\n", "TIME \n", "1906-01-01 29820.0 32.000000\n", "1906-01-02 29820.0 32.000000\n", "1906-01-03 29820.0 27.666667\n", "1906-01-04 29820.0 31.333333\n", "1906-01-05 29820.0 31.666667" ] }, "execution_count": 35, "metadata": {}, "output_type": "execute_result" } ], "source": [ "daily_averages = data.resample(\"1D\").mean()\n", "daily_averages.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As a result we got a new DataFrame, in which all columns having numerical values were aggregated to daily averages. Notice, that averaging all numerical values in your DataFrame might not be what you want, e.g. here it does not really make any sense to take an average of the station number or any other columns expect the `TEMP_F` column. We can easily specify which columns should be aggregated by passing a list of column names after the `resample()` method:" ] }, { "cell_type": "code", "execution_count": 36, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TEMP_F
TIME
1906-01-0132.000000
1906-01-0232.000000
1906-01-0327.666667
1906-01-0431.333333
1906-01-0531.666667
\n", "
" ], "text/plain": [ " TEMP_F\n", "TIME \n", "1906-01-01 32.000000\n", "1906-01-02 32.000000\n", "1906-01-03 27.666667\n", "1906-01-04 31.333333\n", "1906-01-05 31.666667" ] }, "execution_count": 36, "metadata": {}, "output_type": "execute_result" } ], "source": [ "daily_average_temps = data.resample(\"1D\")[[\"TEMP_F\"]].mean()\n", "daily_average_temps.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As a result, we got a new pandas DataFrame containing the average temperature for each day. It is very easy to change the resampling method, and e.g. find out the lowest temperatures for each week using the `.min()` function:" ] }, { "cell_type": "code", "execution_count": 37, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TEMP_F
TIME
1906-01-0726.0
1906-01-1430.0
1906-01-2121.0
1906-01-288.0
1906-02-0418.0
\n", "
" ], "text/plain": [ " TEMP_F\n", "TIME \n", "1906-01-07 26.0\n", "1906-01-14 30.0\n", "1906-01-21 21.0\n", "1906-01-28 8.0\n", "1906-02-04 18.0" ] }, "execution_count": 37, "metadata": {}, "output_type": "execute_result" } ], "source": [ "weekly_min_temps = data.resample(\"1W\")[[\"TEMP_F\"]].min()\n", "weekly_min_temps.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In a quite similar manner, it is possible to do upsampling and e.g. interpolate hourly observations to higher frequency (e.g. every 30 minute) by filling the missing values using e.g. linear interpolation which \"guesses\" the missing values between two known points by fitting a straight line between the observations. Thus, the interpolated temperature between two consecutive hours is the average the two (as discussed in the beginning of this section):" ] }, { "cell_type": "code", "execution_count": 38, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TEMP_F
TIME
2018-01-01 03:00:0034.0
2018-01-01 03:30:0034.5
2018-01-01 04:00:0035.0
2018-01-01 04:30:0035.5
2018-01-01 05:00:0036.0
2018-01-01 05:30:0036.0
2018-01-01 06:00:0036.0
2018-01-01 06:30:0036.5
2018-01-01 07:00:0037.0
2018-01-01 07:30:0037.0
2018-01-01 08:00:0037.0
\n", "
" ], "text/plain": [ " TEMP_F\n", "TIME \n", "2018-01-01 03:00:00 34.0\n", "2018-01-01 03:30:00 34.5\n", "2018-01-01 04:00:00 35.0\n", "2018-01-01 04:30:00 35.5\n", "2018-01-01 05:00:00 36.0\n", "2018-01-01 05:30:00 36.0\n", "2018-01-01 06:00:00 36.0\n", "2018-01-01 06:30:00 36.5\n", "2018-01-01 07:00:00 37.0\n", "2018-01-01 07:30:00 37.0\n", "2018-01-01 08:00:00 37.0" ] }, "execution_count": 38, "metadata": {}, "output_type": "execute_result" } ], "source": [ "selection = data.loc[\"2018-01-01 03:00\":\"2018-01-01 08:00\"].copy()\n", "selection.resample(\"30T\").interpolate()[[\"TEMP_F\"]]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As a result, our hourly data was now converted to higher frequency and e.g. the temperature at 03:30 is 34.5 Fahrenheits which is the average of its neighboring values. By default, the interpolation will automatically done for all columns containing numerical values and you can select the columns that you are interested in by passing a list of column names after calling the `interpolate()` method. There are also many other interpolation methods available, see details from the pandas documentation [^interpolation]. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Rolling statistics** is yet another useful technique to deal with time series data that is \"noisy\" or that has gaps between observations. Moving average is one such an operation in which observations are weighted (averaged) over a sliding \"window\" that is an efficient way to smooth the data and reduce variance between observations. For instance, stock market data or GPS observations representing the movement / speed profiles of vehicles are good examples of datasets that typically have a lot of variance, which can make it challenging to see patterns without smoothing. Pandas supports various moving window functions that have differences in how they influence the end result. Some of them use a static window size and weights the observations equally, whereas some functions can e.g. give more weight to more recent observations. Next we will see how moving average can be calculated with pandas, and how it influences our data. Let's start by taking a sample of our data covering the last 20 years and make a plot based on the temperature:" ] }, { "cell_type": "code", "execution_count": 39, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "sample = data.loc[\"1990\":\"2020\"].copy()\n", "sample.plot(y=\"TEMP_F\");" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "_**Figure 3.5**. Temperature values in Fahrenheit for 1990-2020._\n", "\n", "In this plot we have time represented on the x-axis and the temperature in Fahrenheit on the y-axis. We can clearly see how the temperatures fluctuate over different seasons. Let's investigate our data a bit further and take a look of the first rows of our sample: " ] }, { "cell_type": "code", "execution_count": 40, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
STATION_NUMBERTEMP_F
TIME
1990-01-01 00:00:002982032.0
1990-01-01 03:00:002982031.0
1990-01-01 06:00:002982031.0
1990-01-01 09:00:002982031.0
1990-01-01 12:00:002982031.0
1990-01-01 15:00:002982030.0
1990-01-01 18:00:002982028.0
1990-01-01 21:00:002982029.0
1990-01-02 00:00:002982029.0
1990-01-02 03:00:002982029.0
\n", "
" ], "text/plain": [ " STATION_NUMBER TEMP_F\n", "TIME \n", "1990-01-01 00:00:00 29820 32.0\n", "1990-01-01 03:00:00 29820 31.0\n", "1990-01-01 06:00:00 29820 31.0\n", "1990-01-01 09:00:00 29820 31.0\n", "1990-01-01 12:00:00 29820 31.0\n", "1990-01-01 15:00:00 29820 30.0\n", "1990-01-01 18:00:00 29820 28.0\n", "1990-01-01 21:00:00 29820 29.0\n", "1990-01-02 00:00:00 29820 29.0\n", "1990-01-02 03:00:00 29820 29.0" ] }, "execution_count": 40, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sample.head(10)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As we see, the first rows in our sample contains observations every 3 hours or so. Let's test how we can apply rolling statistic over our tiem series and calculate a moving average based on four closest (consecutive) temperature observations. To do this, we can use a method called `.rolling()` that by default takes an integer number as an argument that specify how many observations should be grouped together (i.e. window size). By chaining the `.mean()` command after the first one, we can specify that we want to calculate the average temperature based on those four observations:" ] }, { "cell_type": "code", "execution_count": 41, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "TIME\n", "1990-01-01 00:00:00 NaN\n", "1990-01-01 03:00:00 NaN\n", "1990-01-01 06:00:00 NaN\n", "1990-01-01 09:00:00 31.25\n", "1990-01-01 12:00:00 31.00\n", " ... \n", "2019-10-01 19:00:00 49.00\n", "2019-10-01 20:00:00 48.50\n", "2019-10-01 21:00:00 47.75\n", "2019-10-01 22:00:00 47.00\n", "2019-10-01 23:00:00 46.75\n", "Name: TEMP_F, Length: 119407, dtype: float64" ] }, "execution_count": 41, "metadata": {}, "output_type": "execute_result" } ], "source": [ "roll_a = sample[\"TEMP_F\"].rolling(4).mean()\n", "roll_a" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "As a result, we get a Series in which the time is the index and the newly calculated average temperatures are the values. As can be seen from here, the first three temperature values are NaNs which is expected, as we specified that the average should always be based on four consecutive observations. Hence, with the first three observations the average cannot be calculated, and in this case pandas returns NaN. \n", "\n", "An important thing to notice here, is that the temporal resolution of our data seem to vary. This can be seen by comparing the first and last five timestamps to each other: in early 90's the frequency of the observations was 3 hours, but during 2019 the frequency has been 1 hour. This is a significant issue, because the temporal span of how the moving average is calculated differs, which can cause unwanted consequenses and errors in our analysis. Luckily, it is easy to fix this issue by passing a *{term}`DateOffset`* as an argument for the `.rolling()` command, in a similar manner as we did earlier when resampling our data. Hence, we can e.g. use a fixed window size of 1 day by passing \"1D\" to the function, and it will always keep the time span identical, no matter how many observations there would be recorded within a single day:" ] }, { "cell_type": "code", "execution_count": 42, "metadata": {}, "outputs": [], "source": [ "daily = sample[\"TEMP_F\"].rolling(\"1D\").mean()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Following this approach, it is easy to calculate rolling statistics with different window sizes, such as weekly, monthly or even annual averages. The main different between `.rolling()` and `.resample()` is that the former will maintain the number of observations in the data (but change the values according neighboring values e.g. based on mean), whereas the latter will actually aggregate and reduce the number of observations that will be kept in the result. To demonstrate the influence of \"smoothing\" to our time series based on moving average, let's calculate the moving average of temperatures based on window size of 180 days and 360 days (~year), and plot the results (details about plotting are discussed in the next chapter):" ] }, { "cell_type": "code", "execution_count": 43, "metadata": {}, "outputs": [], "source": [ "half_year = sample[\"TEMP_F\"].rolling(\"180D\").mean()\n", "annual = sample[\"TEMP_F\"].rolling(\"360D\").mean()" ] }, { "cell_type": "code", "execution_count": 44, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Plot\n", "ax = daily.plot(y=\"TEMP_F\", figsize=(12, 8), alpha=0.5)\n", "ax = half_year.plot(ax=ax, y=\"TEMP_F\", lw=2)\n", "ax = annual.plot(ax=ax, y=\"TEMP_F\", lw=3)\n", "ax.set_ylabel(\"Temperature (Fahrenheit)\", size=12);" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "_**Figure 3.6**. Temperature values and moving averages for 1990-2020._\n", "\n", "As a result, we have three different curves showing how the temperature values get smoothed by averaging. Green line is based on the moving average with 360 days; Orange line is based on the 180 days; and the blue line represents the moving average based on daily averages. As can be seen, the variance between observations is reduced (i.e. smoothed) when the size of the window gets larger. \n", "\n", "In quite a similar manner, it is possible to use exponentially weighted (EW) window functions to give more weight on more recent observations based on a decay function, such as center of mass (com) following a formula: $\\alpha$ *= 1 / (1 + com)*. A method called `.ewm()` does this, and as a result the smoothing adapts faster to changes than using equally weighted observations. The higher value you pass for the `com` -parameter, the more the end result will be smoothed:" ] }, { "cell_type": "code", "execution_count": 45, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "sample[\"TEMP_F\"].ewm(com=200).mean().plot();" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "_**Figure 3.7**. Exponentially weighted temperature values for 1990-2020._\n", "\n", "In addition to these moving window functions, pandas provides a method `.expanding()` which increases the size of the window on each observation until it encompasses the whole series. This method basically cumulates the information on each observation and can be useful e.g. in some predictive applications where cumulative information about the given phenomena can provide more accurate results after more observations have been observed. It is also possible to apply and chain your own functions together with `.rolling()` and `.expanding()`, such as `.rolling(20).apply(myOwnFunction)`. " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Footnotes\n", "\n", "[^isostandard]: \n", "[^dateoffsets]: \n", "[^interpolation]: " ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.9" } }, "nbformat": 4, "nbformat_minor": 4 }