You cannot select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
121 KiB
121 KiB
Лабораторная 1. Интерактивный анализ данных велопарковок SF Bay Area Bike Share в Apache Spark с использованием Spark SQL и DataFrame API¶
Описание данных¶
https://www.kaggle.com/benhamner/sf-bay-area-bike-share
stations.csv схема:
id: station ID number
name: name of station
lat: latitude
long: longitude
dock_count: number of total docks at station
city: city (San Francisco, Redwood City, Palo Alto, Mountain View, San Jose)
installation_date: original date that station was installed. If station was moved, it is noted below.
trips.csv схема:
id: numeric ID of bike trip
duration: time of trip in seconds
start_date: start date of trip with date and time, in PST
start_station_name: station name of start station
start_station_id: numeric reference for start station
end_date: end date of trip with date and time, in PST
end_station_name: station name for end station
end_station_id: numeric reference for end station
bike_id: ID of bike used
subscription_type: Subscriber = annual or 30-day member; Customer = 24-hour or 3-day member
zip_code: Home zip code of subscriber (customers can choose to manually enter zip at kiosk however data is unreliable)
In [1]:
from pyspark import SparkContext, SparkConf from pyspark.sql import SparkSession import pyspark.sql as sql
In [2]:
conf = SparkConf().setAppName("L1_interactive_bike_analysis").setMaster('yarn')
In [3]:
sc = SparkContext(conf=conf) spark = SparkSession(sc)
Пример чтения csv файлов и работы с дефектными данными¶
Список опций чтения и записи для CSV файлов https://spark.apache.org/docs/latest/sql-data-sources-csv.html#data-source-option
Формат паттерна временной метки Spark SQL отличается от python библиотеки datetime. https://spark.apache.org/docs/latest/sql-ref-datetime-pattern.html
In [32]:
tripData = spark.read\ .option("header", True)\ .option("inferSchema", True)\ .option("timestampFormat", 'M/d/y H:m')\ .csv("trips.csv") tripData
Out[32]:
DataFrame[id: int, duration: int, start_date: timestamp, start_station_name: string, start_station_id: int, end_date: timestamp, end_station_name: string, end_station_id: int, bike_id: int, subscription_type: string, zip_code: string]
In [34]:
tripData.printSchema()
root |-- id: integer (nullable = true) |-- duration: integer (nullable = true) |-- start_date: timestamp (nullable = true) |-- start_station_name: string (nullable = true) |-- start_station_id: integer (nullable = true) |-- end_date: timestamp (nullable = true) |-- end_station_name: string (nullable = true) |-- end_station_id: integer (nullable = true) |-- bike_id: integer (nullable = true) |-- subscription_type: string (nullable = true) |-- zip_code: string (nullable = true)
In [35]:
tripData.show(n=5)
+----+--------+-------------------+--------------------+----------------+-------------------+--------------------+--------------+-------+-----------------+--------+ | id|duration| start_date| start_station_name|start_station_id| end_date| end_station_name|end_station_id|bike_id|subscription_type|zip_code| +----+--------+-------------------+--------------------+----------------+-------------------+--------------------+--------------+-------+-----------------+--------+ |4576| 63| null|South Van Ness at...| 66|2013-08-29 14:14:00|South Van Ness at...| 66| 520| Subscriber| 94127| |4607| null|2013-08-29 14:42:00| San Jose City Hall| 10|2013-08-29 14:43:00| San Jose City Hall| 10| 661| Subscriber| 95138| |4130| 71|2013-08-29 10:16:00|Mountain View Cit...| 27|2013-08-29 10:17:00|Mountain View Cit...| 27| 48| Subscriber| 97214| |4251| 77|2013-08-29 11:29:00| San Jose City Hall| 10|2013-08-29 11:30:00| San Jose City Hall| 10| 26| Subscriber| 95060| |4299| 83|2013-08-29 12:02:00|South Van Ness at...| 66|2013-08-29 12:04:00| Market at 10th| 67| 319| Subscriber| 94103| +----+--------+-------------------+--------------------+----------------+-------------------+--------------------+--------------+-------+-----------------+--------+ only showing top 5 rows
In [26]:
? tripData.dropna
Signature: tripData.dropna(how='any', thresh=None, subset=None) Docstring: Returns a new :class:`DataFrame` omitting rows with null values. :func:`DataFrame.dropna` and :func:`DataFrameNaFunctions.drop` are aliases of each other. .. versionadded:: 1.3.1 Parameters ---------- how : str, optional 'any' or 'all'. If 'any', drop a row if it contains any nulls. If 'all', drop a row only if all its values are null. thresh: int, optional default None If specified, drop rows that have less than `thresh` non-null values. This overwrites the `how` parameter. subset : str, tuple or list, optional optional list of column names to consider. Examples -------- >>> df4.na.drop().show() +---+------+-----+ |age|height| name| +---+------+-----+ | 10| 80|Alice| +---+------+-----+ File: ~/.local/lib/python3.9/site-packages/pyspark/sql/dataframe.py Type: method
In [36]:
tripData.dropna().show(n=5)
+----+--------+-------------------+--------------------+----------------+-------------------+--------------------+--------------+-------+-----------------+--------+ | id|duration| start_date| start_station_name|start_station_id| end_date| end_station_name|end_station_id|bike_id|subscription_type|zip_code| +----+--------+-------------------+--------------------+----------------+-------------------+--------------------+--------------+-------+-----------------+--------+ |4130| 71|2013-08-29 10:16:00|Mountain View Cit...| 27|2013-08-29 10:17:00|Mountain View Cit...| 27| 48| Subscriber| 97214| |4251| 77|2013-08-29 11:29:00| San Jose City Hall| 10|2013-08-29 11:30:00| San Jose City Hall| 10| 26| Subscriber| 95060| |4299| 83|2013-08-29 12:02:00|South Van Ness at...| 66|2013-08-29 12:04:00| Market at 10th| 67| 319| Subscriber| 94103| |4927| 103|2013-08-29 18:54:00| Golden Gate at Polk| 59|2013-08-29 18:56:00| Golden Gate at Polk| 59| 527| Subscriber| 94109| |4500| 109|2013-08-29 13:25:00|Santa Clara at Al...| 4|2013-08-29 13:27:00| Adobe on Almaden| 5| 679| Subscriber| 95112| +----+--------+-------------------+--------------------+----------------+-------------------+--------------------+--------------+-------+-----------------+--------+ only showing top 5 rows
In [37]:
tripData.describe().show()
+-------+------------------+------------------+--------------------+------------------+--------------------+------------------+------------------+-----------------+-------------------+ |summary| id| duration| start_station_name| start_station_id| end_station_name| end_station_id| bike_id|subscription_type| zip_code| +-------+------------------+------------------+--------------------+------------------+--------------------+------------------+------------------+-----------------+-------------------+ | count| 985352| 985351| 985352| 985352| 985352| 985352| 985352| 985352| 978484| | mean|521401.35102481145|1092.1337716204682| null|58.060558054380564| null|58.043319544690625|426.07050779822845| null| 2008421.845039644| | stddev|245889.28553182338| 25689.80328664839| null|16.998928220336474| null| 17.10456379871572| 155.4614377036669| null|1.243190730418017E9| | min| 4069| 60| 2nd at Folsom| 2| 2nd at Folsom| 2| 9| Customer| 0| | max| 913460| 17270400|Yerba Buena Cente...| 84|Yerba Buena Cente...| 84| 878| Subscriber| v6z2x| +-------+------------------+------------------+--------------------+------------------+--------------------+------------------+------------------+-----------------+-------------------+
In [39]:
stationData = spark.read\ .option("header", True)\ .option("inferSchema", True)\ .option("timestampFormat", 'M/d/y')\ .csv("stations.csv") stationData.printSchema()
root |-- id: integer (nullable = true) |-- name: string (nullable = true) |-- lat: double (nullable = true) |-- long: double (nullable = true) |-- dock_count: integer (nullable = true) |-- city: string (nullable = true) |-- installation_date: timestamp (nullable = true)
In [40]:
stationData.show(n=5)
+---+--------------------+------------------+-------------------+----------+--------+-------------------+ | id| name| lat| long|dock_count| city| installation_date| +---+--------------------+------------------+-------------------+----------+--------+-------------------+ | 2|San Jose Diridon ...| 37.329732|-121.90178200000001| 27|San Jose|2013-08-06 00:00:00| | 3|San Jose Civic Ce...| 37.330698| -121.888979| 15|San Jose|2013-08-05 00:00:00| | 4|Santa Clara at Al...| 37.333988| -121.894902| 11|San Jose|2013-08-06 00:00:00| | 5| Adobe on Almaden| 37.331415| -121.8932| 19|San Jose|2013-08-05 00:00:00| | 6| San Pedro Square|37.336721000000004| -121.894074| 15|San Jose|2013-08-07 00:00:00| +---+--------------------+------------------+-------------------+----------+--------+-------------------+ only showing top 5 rows
In [41]:
stationData.describe().show()
+-------+------------------+--------------------+-------------------+-------------------+-----------------+-------------+ |summary| id| name| lat| long| dock_count| city| +-------+------------------+--------------------+-------------------+-------------------+-----------------+-------------+ | count| 70| 70| 70| 70| 70| 70| | mean| 43.0| null| 37.59024338428572|-122.21841616428571|17.65714285714286| null| | stddev|24.166091947189145| null|0.20347253639672502|0.20944604979644524|4.010441857493954| null| | min| 2| 2nd at Folsom| 37.329732| -122.418954| 11|Mountain View| | max| 84|Yerba Buena Cente...| 37.80477| -121.877349| 27| San Jose| +-------+------------------+--------------------+-------------------+-------------------+-----------------+-------------+
Пример использования DataFrame API¶
Выполните операцию объединения коллекций по ключу с помощью функции join. Объедините stationsIndexed и tripsByStartTerminals, stationsIndexed и tripsByEndTerminals.
In [42]:
tripData.printSchema() stationData.printSchema()
root |-- id: integer (nullable = true) |-- duration: integer (nullable = true) |-- start_date: timestamp (nullable = true) |-- start_station_name: string (nullable = true) |-- start_station_id: integer (nullable = true) |-- end_date: timestamp (nullable = true) |-- end_station_name: string (nullable = true) |-- end_station_id: integer (nullable = true) |-- bike_id: integer (nullable = true) |-- subscription_type: string (nullable = true) |-- zip_code: string (nullable = true) root |-- id: integer (nullable = true) |-- name: string (nullable = true) |-- lat: double (nullable = true) |-- long: double (nullable = true) |-- dock_count: integer (nullable = true) |-- city: string (nullable = true) |-- installation_date: timestamp (nullable = true)
In [45]:
stationsView = stationData.select(stationData['id'], stationData['name'], stationData['lat'], stationData['long']) stationsView.show()
+---+--------------------+------------------+-------------------+ | id| name| lat| long| +---+--------------------+------------------+-------------------+ | 2|San Jose Diridon ...| 37.329732|-121.90178200000001| | 3|San Jose Civic Ce...| 37.330698| -121.888979| | 4|Santa Clara at Al...| 37.333988| -121.894902| | 5| Adobe on Almaden| 37.331415| -121.8932| | 6| San Pedro Square|37.336721000000004| -121.894074| | 7|Paseo de San Antonio| 37.333798|-121.88694299999999| | 8| San Salvador at 1st| 37.330165|-121.88583100000001| | 9| Japantown| 37.348742|-121.89471499999999| | 10| San Jose City Hall| 37.337391| -121.886995| | 11| MLK Library| 37.335885|-121.88566000000002| | 12|SJSU 4th at San C...| 37.332808|-121.88389099999999| | 13| St James Park| 37.339301|-121.88993700000002| | 14|Arena Green / SAP...| 37.332692| -121.900084| | 16|SJSU - San Salvad...|37.333954999999996| -121.877349| | 21| Franklin at Maple| 37.481758| -122.226904| | 22|Redwood City Calt...|37.486078000000006|-122.23208899999999| | 23|San Mateo County ...|37.487615999999996| -122.229951| | 24|Redwood City Publ...| 37.484219| -122.227424| | 25|Stanford in Redwo...| 37.48537|-122.20328799999999| | 26|Redwood City Medi...| 37.487682| -122.223492| +---+--------------------+------------------+-------------------+ only showing top 20 rows
In [56]:
startTrips = tripData.select(tripData.id, tripData.duration, tripData.start_station_id).withColumnRenamed('id', 'trip_id').join(stationsView, tripData.start_station_id == stationsView.id) startTrips = startTrips.drop('id')
In [57]:
startTrips.show()
+-------+--------+----------------+--------------------+------------------+-------------------+ |trip_id|duration|start_station_id| name| lat| long| +-------+--------+----------------+--------------------+------------------+-------------------+ | 4576| 63| 66|South Van Ness at...| 37.774814| -122.418954| | 4607| null| 10| San Jose City Hall| 37.337391| -121.886995| | 4130| 71| 27|Mountain View Cit...| 37.389218| -122.081896| | 4251| 77| 10| San Jose City Hall| 37.337391| -121.886995| | 4299| 83| 66|South Van Ness at...| 37.774814| -122.418954| | 4927| 103| 59| Golden Gate at Polk| 37.781332| -122.418603| | 4500| 109| 4|Santa Clara at Al...| 37.333988| -121.894902| | 4563| 111| 8| San Salvador at 1st| 37.330165|-121.88583100000001| | 4760| 113| 66|South Van Ness at...| 37.774814| -122.418954| | 4258| 114| 10| San Jose City Hall| 37.337391| -121.886995| | 4549| 125| 49| Spear at Folsom|37.790302000000004|-122.39063700000001| | 4498| 126| 6| San Pedro Square|37.336721000000004| -121.894074| | 4965| 129| 28|Mountain View Cal...|37.394358000000004|-122.07671299999998| | 4557| 130| 64| 2nd at South Park| 37.782259| -122.392738| | 4386| 134| 41| Clay at Battery| 37.795001| -122.39997| | 4749| 138| 47| Post at Kearney| 37.788975| -122.403452| | 4242| 141| 10| San Jose City Hall| 37.337391| -121.886995| | 4329| 142| 67| Market at 10th|37.776619000000004|-122.41738500000001| | 5097| 142| 74| Steuart at Market| 37.794139| -122.394434| | 5084| 144| 39| Powell Street BART|37.783871000000005| -122.408433| +-------+--------+----------------+--------------------+------------------+-------------------+ only showing top 20 rows
In [63]:
tripData.printSchema() stationData.printSchema()
root |-- id: integer (nullable = true) |-- duration: integer (nullable = true) |-- start_date: timestamp (nullable = true) |-- start_station_name: string (nullable = true) |-- start_station_id: integer (nullable = true) |-- end_date: timestamp (nullable = true) |-- end_station_name: string (nullable = true) |-- end_station_id: integer (nullable = true) |-- bike_id: integer (nullable = true) |-- subscription_type: string (nullable = true) |-- zip_code: string (nullable = true) root |-- id: integer (nullable = true) |-- name: string (nullable = true) |-- lat: double (nullable = true) |-- long: double (nullable = true) |-- dock_count: integer (nullable = true) |-- city: string (nullable = true) |-- installation_date: timestamp (nullable = true)
Пример использования Spark SQL API¶
In [65]:
stationData.createOrReplaceTempView("stations") tripData.createOrReplaceTempView("trips")
In [74]:
endTrips = spark.sql(""" SELECT trips.id as trip_id, trips.end_station_id, trips.duration, stations.name as station_name, stations.lat, stations.long FROM trips INNER JOIN stations ON trips.end_station_id==stations.id """)
In [75]:
endTrips.show()
+-------+--------------+--------+--------------------+------------------+-------------------+ |trip_id|end_station_id|duration| station_name| lat| long| +-------+--------------+--------+--------------------+------------------+-------------------+ | 4576| 66| 63|South Van Ness at...| 37.774814| -122.418954| | 4607| 10| null| San Jose City Hall| 37.337391| -121.886995| | 4130| 27| 71|Mountain View Cit...| 37.389218| -122.081896| | 4251| 10| 77| San Jose City Hall| 37.337391| -121.886995| | 4299| 67| 83| Market at 10th|37.776619000000004|-122.41738500000001| | 4927| 59| 103| Golden Gate at Polk| 37.781332| -122.418603| | 4500| 5| 109| Adobe on Almaden| 37.331415| -121.8932| | 4563| 8| 111| San Salvador at 1st| 37.330165|-121.88583100000001| | 4760| 66| 113|South Van Ness at...| 37.774814| -122.418954| | 4258| 11| 114| MLK Library| 37.335885|-121.88566000000002| | 4549| 54| 125|Embarcadero at Br...| 37.787152|-122.38801299999999| | 4498| 4| 126|Santa Clara at Al...| 37.333988| -121.894902| | 4965| 28| 129|Mountain View Cal...|37.394358000000004|-122.07671299999998| | 4557| 64| 130| 2nd at South Park| 37.782259| -122.392738| | 4386| 56| 134| Beale at Market| 37.792251|-122.39708600000002| | 4749| 47| 138| Post at Kearney| 37.788975| -122.403452| | 4242| 10| 141| San Jose City Hall| 37.337391| -121.886995| | 4329| 67| 142| Market at 10th|37.776619000000004|-122.41738500000001| | 5097| 50| 142|Harry Bridges Pla...| 37.795392| -122.394203| | 5084| 76| 144| Market at 4th| 37.786305|-122.40496599999999| +-------+--------------+--------+--------------------+------------------+-------------------+ only showing top 20 rows
Для каждой стартовой станции найдем среднее время поездки.
Рассчитаем среднее время поездки для каждого стартового парковочного места
In [225]:
spark.sql(""" SELECT start_station_name, avg(duration) FROM trips GROUP BY trips.start_station_name ORDER BY avg(duration) DESC """).show()
+--------------------+------------------+ | start_station_name| avg(duration)| +--------------------+------------------+ |University and Em...| 7230.231034482758| |Redwood City Medi...| 4764.68287037037| |San Jose Civic Ce...| 4720.621422376409| | Park at Olive| 4686.397612488521| |California Ave Ca...| 4502.620639534884| |Redwood City Publ...|3697.0892307692307| |Rengstorff Avenue...| 3544.797270955166| |Palo Alto Caltrai...| 3158.331498866947| |San Mateo County ...|3002.0827067669175| |South Van Ness at...|2936.8873503613395| |San Antonio Shopp...| 2508.434736091298| |Cowper at University| 2493.220572640509| | Broadway at Main|2481.2537313432836| |Redwood City Calt...| 2405.29409190372| | Japantown|2297.0913838120105| |San Antonio Caltr...|2103.7238932071646| |SJSU 4th at San C...| 1995.366021236727| |Washington at Kea...|1979.3077445652175| | Mezes Park|1918.1354359925788| |Arena Green / SAP...|1888.3390476190475| +--------------------+------------------+ only showing top 20 rows
Пример подготовки данных c Spark SQL, pandas, h3 для их визуализации на карте folium¶
In [ ]:
# ! pip install h3 h3_pyspark pandas folium
Найдём велосипеды, которые ездили в рождество 2014 года. https://spark.apache.org/docs/latest/api/sql/#make_timestamp
In [86]:
# year - the year to represent, from 1 to 9999 # month - the month-of-year to represent, from 1 (January) to 12 (December) # day - the day-of-month to represent, from 1 to 31 # hour - the hour-of-day to represent, from 0 to 23 # min - the minute-of-hour to represent, from 0 to 59 # sec - the second-of-minute and its micro-fraction to represent, from 0 to 60. The value can be either an integer like 13 , or a fraction like 13.123. If the sec argument equals to 60, the seconds field is set to 0 and 1 minute is added to the final timestamp. # timezone - the time zone identifier. For example, CET, UTC and etc. spark.sql(""" SELECT bike_id, start_date, end_date FROM trips WHERE start_date > make_timestamp(2014, 12, 25, 0, 0, 0) AND start_date < make_timestamp(2014, 12, 26, 0, 0, 0) """).show()
+-------+-------------------+-------------------+ |bike_id| start_date| end_date| +-------+-------------------+-------------------+ | 379|2014-12-25 22:10:00|2014-12-25 22:18:00| | 709|2014-12-25 21:21:00|2014-12-25 21:27:00| | 376|2014-12-25 20:40:00|2014-12-25 20:46:00| | 541|2014-12-25 20:27:00|2014-12-25 20:32:00| | 283|2014-12-25 19:56:00|2014-12-25 20:01:00| | 519|2014-12-25 19:56:00|2014-12-25 20:01:00| | 583|2014-12-25 19:05:00|2014-12-25 19:07:00| | 495|2014-12-25 18:42:00|2014-12-25 18:44:00| | 541|2014-12-25 18:28:00|2014-12-25 18:37:00| | 585|2014-12-25 18:27:00|2014-12-25 18:37:00| | 574|2014-12-25 18:12:00|2014-12-25 18:21:00| | 630|2014-12-25 18:12:00|2014-12-25 18:22:00| | 583|2014-12-25 18:05:00|2014-12-25 18:22:00| | 290|2014-12-25 18:01:00|2014-12-25 18:15:00| | 451|2014-12-25 17:55:00|2014-12-25 18:04:00| | 630|2014-12-25 17:55:00|2014-12-25 17:59:00| | 574|2014-12-25 17:54:00|2014-12-25 17:59:00| | 463|2014-12-25 17:46:00|2014-12-25 17:53:00| | 628|2014-12-25 17:46:00|2014-12-25 17:53:00| | 58|2014-12-25 16:32:00|2014-12-25 17:04:00| +-------+-------------------+-------------------+ only showing top 20 rows
Найдём станции через которые проехал один из велосипедов, найденных ранее.
In [93]:
spark.sql(""" SELECT trips.bike_id, trips.start_date, trips.end_date, stations.name FROM trips INNER JOIN stations ON trips.start_station_id == stations.id WHERE bike_id == 583 AND start_date > make_timestamp(2014, 12, 25, 0, 0, 0) AND start_date < make_timestamp(2014, 12, 26, 0, 0, 0) """).show()
+-------+-------------------+-------------------+--------------+ |bike_id| start_date| end_date| name| +-------+-------------------+-------------------+--------------+ | 583|2014-12-25 19:05:00|2014-12-25 19:07:00|Market at 10th| | 583|2014-12-25 18:05:00|2014-12-25 18:22:00|Market at 10th| | 583|2014-12-25 12:14:00|2014-12-25 12:21:00| Market at 4th| | 583|2014-12-25 19:05:00|2014-12-25 19:07:00|Market at 10th| | 583|2014-12-25 18:05:00|2014-12-25 18:22:00|Market at 10th| | 583|2014-12-25 12:14:00|2014-12-25 12:21:00| Market at 4th| +-------+-------------------+-------------------+--------------+
Найдём все станции, которые попали в ту же клетку h3 координатной сетки что и станции, через которые проехал велосипед 583 25.12.2014.
Отобразим gps координаты станций в координаты h3.
In [96]:
from pyspark.sql import functions as F import h3_pyspark import h3
H3 Grid Resolutions https://h3geo.org/docs/core-library/restable/
In [205]:
resolution = 8 stationData.withColumn('h3', h3_pyspark.geo_to_h3('lat', 'long', sql.functions.lit(resolution))).createOrReplaceTempView("stations_h3")
Используя вложенный sql запрос, найдём h3 координаты станций, через который проехал велосипед 583. А затем отфильтруем поездки в рождество 2014 года, которые стартовали со станций с теми же h3 координатами, что мы нашли.
In [166]:
christmas_583_contacts = spark.sql(""" SELECT trips.bike_id, trips.start_date, stations_h3.h3, stations_h3.lat, stations_h3.long, stations_h3.name FROM trips INNER JOIN stations_h3 ON trips.start_station_id == stations_h3.id WHERE stations_h3.h3 in (SELECT stations_h3.h3 FROM trips INNER JOIN stations_h3 ON trips.start_station_id == stations_h3.id WHERE bike_id == 583 AND start_date > make_timestamp(2014, 12, 25, 0, 0, 0) AND start_date < make_timestamp(2014, 12, 26, 0, 0, 0)) AND start_date > make_timestamp(2014, 12, 25, 0, 0, 0) AND start_date < make_timestamp(2014, 12, 26, 0, 0, 0) ORDER BY trips.start_date """) christmas_583_contacts.cache() christmas_583_contacts.show()
+-------+-------------------+---------------+------------------+-------------------+--------------+ |bike_id| start_date| h3| lat| long| name| +-------+-------------------+---------------+------------------+-------------------+--------------+ | 439|2014-12-25 01:40:00|8828308281fffff|37.776619000000004|-122.41738500000001|Market at 10th| | 439|2014-12-25 01:40:00|8828308281fffff|37.776619000000004|-122.41738500000001|Market at 10th| | 659|2014-12-25 09:49:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 465|2014-12-25 09:49:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 659|2014-12-25 09:49:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 465|2014-12-25 09:49:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 583|2014-12-25 12:14:00|88283082abfffff| 37.786305|-122.40496599999999| Market at 4th| | 583|2014-12-25 12:14:00|88283082abfffff| 37.786305|-122.40496599999999| Market at 4th| | 479|2014-12-25 12:22:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 331|2014-12-25 12:22:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 479|2014-12-25 12:22:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 331|2014-12-25 12:22:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 330|2014-12-25 12:27:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 438|2014-12-25 12:27:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 330|2014-12-25 12:27:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 438|2014-12-25 12:27:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 428|2014-12-25 12:35:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 428|2014-12-25 12:35:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 214|2014-12-25 12:38:00|88283082abfffff|37.781752000000004|-122.40512700000001| 5th at Howard| | 292|2014-12-25 12:38:00|88283082abfffff| 37.786305|-122.40496599999999| Market at 4th| +-------+-------------------+---------------+------------------+-------------------+--------------+ only showing top 20 rows
In [167]:
import pandas as pd import h3 h3_places = christmas_583_contacts.select('lat','long', 'name', 'h3').toPandas()
In [215]:
# source code from https://nbviewer.org/github/uber/h3-py-notebooks/blob/master/notebooks/usage.ipynb import folium def init_map(hexagons, width=1100, height=900): lats = [] longs = [] for hexagon in hexagons: lat, long = h3.h3_to_geo(hexagon) lats.append(lat) longs.append(long) return folium.Map(location=[sum(lats)/len(lats), sum(longs)/len(longs)], zoom_start=15, tiles='cartodbpositron', width=width, height=height) def visualize_hexagons(folium_map, hexagons, color="red"): """ hexagons is a list of hexcluster. Each hexcluster is a list of hexagons. eg. [[hex1, hex2], [hex3, hex4]] """ polylines = [] lat = [] lng = [] for hex in hexagons: polygons = h3.h3_set_to_multi_polygon([hex], geo_json=False) # flatten polygons into loops. outlines = [loop for polygon in polygons for loop in polygon] polyline = [outline + [outline[0]] for outline in outlines][0] lat.extend(map(lambda v:v[0], polyline)) lng.extend(map(lambda v:v[1], polyline)) polylines.append(polyline) for polyline in polylines: my_PolyLine = folium.PolyLine(locations=polyline, weight=8, color=color) folium_map.add_child(my_PolyLine) return folium_map def visualize_stations(folium_map, stations, color="red"): """ stations is a dataframe with columns: lat, long, station_name """ for idx, lat, long, station_name in stations.itertuples(): folium_map.add_child(folium.map.Marker(location=(lat, long))) folium_map.add_child(folium.map.Marker(location=(lat, long), icon=folium.features.DivIcon( icon_size=(500,36), icon_anchor=(-17,37), html=f'<div style="display: inline-block;font-size: 10pt; background: rgba(255, 255, 255, 0.8)">{station_name}</div>', ))) return folium_map
In [217]:
m = init_map(h3_places.h3.unique()) visualize_hexagons(m, h3_places.h3.unique(), color="black") visualize_stations(m, h3_places.loc[:, ['lat', 'long', 'name']]) display(m)
Make this Notebook Trusted to load map: File -> Trust Notebook
In [226]:
sc.stop()