Autolib-Project

Independent Project w4 Data Preparation Data was prepared for use using Python. I used the whole dataset and some of the columns were dropped as they did not add any value to the analysis. The following were undertaken when cleaning the data: Dropping unnecessary columns: #dropping the rows that are unnecessary df.drop(['Address','Cars','Charge Slots','Charging Status','Displayed comment','Scheduled at', 'Slots','Geo point'], axis = 1, inplace = True) The remaining columns were: Index(['Bluecar counter', 'Utilib counter', 'Utilib 1.4 counter', 'City', 'ID', 'Kind', 'Postal code', 'Public name', 'Rental status', 'Station type', 'Status', 'Subscription status', 'year', 'month', 'day', 'hour', 'minute'], dtype='object') The columns that remained were all converted to lower case format for ease of referencing while doing the analysis using the following code: #converting column names to lower case df.columns = map(str.lower, df.columns) All column names that has spaces between were all replaced with an underscore using the following code for uniformity among the column names. df.columns = df.columns.str.strip().str.replace(' ','_') The year, month and day columns were then combined to create a new column date that contained all the date information. parking['date'] = pd.to_datetime((parking.year10000+parking.month100+parking.day).apply(str),format='%Y%m%d') The hour and minute columns were also combined to create a new column time. parking['time'] = parking.hour.str.cat([parking.minute], sep = ':') A new Csv was then created ‘cars.csv’ that contained the cleaned dataset that aided in working with a sizeable data with valuable information. Data Analysis In order to get the most popular hour of the day in the city of Paris for picking up the shared Bluecars, I selected only the column where city was Paris in order to do the analysis by subsetting the dataframe. #subsetting the data frame to only select only stations in Paris pariscity = cars[(cars.city == 'Paris')]

bluecar_counter	utilib_counter	utilib_1.4_counter	city	id	kind	postal_code	public_name	rental_status	station_type	status	subscription_status	day	hour	minute	date	time

0 3 1 0 Paris paris-vulpian-24 STATION 75013 Paris/Vulpian/24 operational station ok nonexistent 1 0 0 2018-04-01 0:0 1 0 0 0 Paris paris-richardlenoir-8 STATION 75011 Paris/Richard Lenoir/8 operational station ok nonexistent 1 0 0 2018-04-01 0:0 2 3 0 0 Paris paris-charlesbossut-4 STATION 75012 Paris/Charles Bossut/4 operational station ok nonexistent 1 0 0 2018-04-01 0:0 3 4 0 0 Paris paris-claudevellefaux-3 STATION 75010 Paris/Claude Vellefaux/3 operational station ok nonexistent 1 0 0 2018-04-01 0:0

To get the most popular hour in the City of Paris the following code was used: #most popular hour of the day for picking shared electric blue car in Paris pariscity.groupby(['hour'])['bluecar_counter'].sum() #print(x) #.sort_values(ascending = False)

The most popular hour was got by getting the difference between the hours and most popular was hour 23 which had the most negative number of cars. x.diff().sort_values(ascending = False) 8 -47931.0 23 -54699.0

The most popular hour for returning cars was Hour 21 which had the most positive number of cars: 21 66986.0 2 64544.0

This is also shown in the bar graph below:

The most popular station in Paris was:

pariscity.groupby(['id'])['Car_Total'].sum().sort_values(ascending = False).head() id paris-portesaintcloud-parking 71800 paris-versaillesreynaud-parking 67970 paris-picpusnation-parking 64766 paris-gouvionmaillot-parking 63415 paris-lacordaire-56 60294

The most popular station and most picking hour in Paris in terms of the total number of cars pariscity.groupby(['id','hour'])['Car_Total'].sum().sort_values(ascending = False).head(15) id hour paris-portesaintcloud-parking 12 3238 11 3236 14 3216 18 3213 7 3196 19 3194 10 3190 13 3178 15 3172 16 3162 6 3159 17 3141 paris-versaillesreynaud-parking 5 3106 2 3092 4 3090

The most popular postal code for picking up blue cars cars.groupby(['postal_code','hour'])['bluecar_counter'].sum().sort_values(ascending = False).head()

postal_code hour 75015 6 92407 5 91494 4 89615 3 89528 7 89175 Name: bluecar_counter, dtype: int64

The most popular station for the company was: pariscity.groupby(['id','postal_code'])['Car_Total'].sum().sort_values(ascending = False).head() id postal_code paris-portesaintcloud-parking 75016 71800 paris-versaillesreynaud-parking 75016 67970 paris-picpusnation-parking 75012 64766 paris-gouvionmaillot-parking 75017 63415 paris-lacordaire-56 75015 60294

The most popular station, Paris – Portesaintcloud- parking which is postal code 75016 does not belong to the most popular postal code which is 75015. They differ. The results change when one uses Utilib and Utilib 1.4. With Utilib, the popular station is colombes-gounoud-1 of postal code 92700. id postal_code hour colombes-gounod-1 92700 11 464 10 459 12 447 nanterre-julesquentin-33 92000 17 429 colombes-gounod-1 92700 13 429

With Utilib 1.4, the popular station is sevres-granderue-123 of postal code 92310. id postal_code hour sevres-granderue-123 92310 7 546 drancy-rogersalengro-3 93700 11 545 6 545 2 545 sevres-granderue-123 92310 3 544 Name: utilib_1.4_counter, dtype: int64