Capstone week4

Exploring downtown Toronto Neighborhood with 

the aim of setting up a Supermarket.

               A capstone report by Oladapo Ajayi

Introduction

In setting up a supermarket, what neighborhood(s) are best considered for optimum profit in Downtown Toronto area?

A little background shows that Downtown Toronto is the business hub of the district of Toronto Ontario, Canada. 
Its land mass is 14 square kilometer. It is in between Bloor street, which is to the north of the city Don Valley and Bathurst to the the east 
and west respectively. It’s a very busy area as this is also attributed to the place being the government residential 
area apart from the fact that it is a major business area.

Based on the aforementioned, I shall focus this report on DownTown Toronto area as it will be a perfect area to 
perform this research as a result of its unique characteristics. Ti have more information on this area, kindly visit the 
Wikipedia page: https://en.wikipedia.org/wiki/Downtown_Toronto 

Data Source

I will be using the Foursquare API

Due to the richness and available from the Foursquare API, the data from this source shall be used for this research. 
It is a a good place to get excellent API for developers. Others include google. I will be using Foursquare because it is very rich and detailed.

Fetching Latitude and Longitude

I will be use a geocoder to get the latitude and longitude coordinates for the DownTown Toronto.

Web Scrapping

In order to get the important data for this research, I shall be scrapping the Wikipedia page so has to have the list of boroughs and 
neighborhoods in Toronto. This is particularly important for my analysis.

I will be using the Wikipedia page below

https://en.wikipedia.org/wiki/List_of_postal_codes_of_Canada: 

Methodology

1. Downloading and cleaning data

In this step we scrape the required data from the wikipedia page given above. We clean the table and 
extract only the information that is needed into a pandas dataframe. i.e. for Toronto City

A detailed process of downloading and extracting 
data can be found in the link: here

2. Geocoding the locations to obtain georaphical coordinates (latitude, longitude)

Now that you have built a dataframe of the postal code of each neighborhood along with the borough name 
and neighborhood name, in order to utilize the Foursquare location data, we need to get the latitude and the longitude coordinates 
of each neighborhood. We use geocoder library to accomplish this. However, given that this package can be very unreliable, 
in case we are not able to get the geographical coordinates of the neighborhoods using the Geocoder package, 
here is a link to a csv file that has the geographical coordinates of each postal code: http://cocl.us/Geospatial_data

A detailed process of geocoding can be found in the below link:

https://github.com/Oladapo1986/My-Capstone-Project/blob/master/geo%20coder.ipynb

3. Explore and Analyze the Downtown Toronto Neighbourhood Data

Here we use the Foursquare API to explore neighborhoods in Toronto City. We will use the explore function to get the 
most common venue categories in each neighborhood, and then use this feature to group the neighborhoods into clusters.

Machine Learning

I will use the k-means clustering algorithm to complete this task. Finally, we will use the 
Folium library to visualize the neighborhoods in Downtown Toronto and their emerging clusters.

Results and Discussion of Findings

Given that a successful data has been obtained on Downtown Toronto data, then data analysis can be carried 
out in order to aid the results and discussion of findings.

The below graph shows distribution by histograph:  Rosedale, Christie and CN Tower/Island Airport area have some the least 
venues in terms of numbers assuming venues cannot be greater than 100 for each area plotted below It can be observed that 
there are over 100 coffee shops, over 80 cafes and nearly 45 restaurants & hotels in Downtown Toronto.

We also find that there are only around 10 supermarkets as shown in the last 5th bar in the map in the whole of Downtown Toronto.

Here is a table showing the top 10 most common venues for each neighborhood in Downtown Toronto.

It is observed that a supermarket is not in anyway visible e in any top 10 venues of any neighborhood.

The next step is to cluster these neighborhoods based on the above info and plot them on the map to see the overall 
emerging clusters.

Cluster 0 (Colored red)

This cluster has all the right mix. 

This cluster is a perfect mixture for different venue categories. There are a lots of restaurants and fast food joints. It also has many 

options for recreational activities like pubs, nightclubs and parks.

Cluster 1 (Colored purple)

There is an airport here. It may not be advisable to set up here

Cluster 2 (Colored blue)

We can observe that eateries and men shops are common here

Cluster 3 (Green)

The cluster in this regard has similar characteristics with clusters under zero.

Cluster 4 (Orange)

Berczy Park could be the best place to set up the supermarket.

Conclusion

In this report we have examined the Downtown Toronto data to find out which would be an ideal location to setup a grocery store. 
We have explored the venue data obtained from Foursquare API and the data obtained by scraping the web. 
We have then applied various plotting and machine learning techniques to further explore the insights. 
We find that grocery store are not as common as coffee shops or restaurants. 
There are only nearly 10 grocery store in Downtown Toronto Neighborhood. This study can be used to decide 
whether a location is suitable for setting up a grocery store or not.

Future Scope

But still there is a LARGE scope of improvement in the dataset. We did not consider one main data entity in this analysis which is People.

A very crucial point while starting up a business is to analyze the neighborhoods