2. FINDING BEST PLACE TO START A
RESTAURANT IN TORONTO
We have explored New York City and the city of Toronto and segmented and
clustered their neighbourhoods. Both cities are very diverse and are the financial
capitals of their respective countries.
The purpose of this project is to help them better understand the Geographical
advantages of different districts so that they can have a best choice of all kinds of
investments.
In addition, the factors on other advantages and disadvantages are considered.
The main audience for this session would be business person who is willing to
start a restaurant around the city.
3. DATA GATHERING AND CLEANING
City Postal Codes, Borough, Neighbourhood have been captured from Wikipedia
page https://en.Wikipedia.org/wiki/List_of_postal_codes_of_Canada:_M
City Latitudes, Longitudes have been captured from the CSV file readily available
@ http://cocl.us/Geospatial_data
Postal Codes with not assigned borough have been re-assigned with their
neighbourhood names
Two DFs have same foreign value as Postal codes.
4. MERGE OF TWO DFs
While merging two DFs using the common value Postal codes, we could able to
see the outcome as follows:
5. FOLIUM MAP DATA
Folium Map API helped us here to show
the location coordinate with blue colour.
All the distinct locations have been
separated by this highlighted blue colour
circle points.
We could able to zoom in and find the
exact details in folium maps unless in
other maps API.
Maps showing City of Toronto, Canada.
6. FOURSQUARE API
The Foursquare API provides location based experiences with diverse information
about venues, users, photos, and check-ins. The API supports real time access to
places, Snap-to-Place that assigns users to specific locations, and Geo-tag.
Additionally, Foursquare allows developers to build audience segments for
analysis and measurement.
We will be given an unique client ID and secret to make use of the data available
from Foursquare API.
We are getting some of the pre-defined functions from Foursquare API Developer
page to facilitate our need and produce exact results as we require.
7. FOURSQUARE API Contd.
As we process the location coordinates in to Foursquare API and it will produce
the output as follows.
So it is giving us the venue based on the radius we are providing for the API call
and its corresponding frequency rate.
8. K – MEANS CLUSTERING
We are using one of the classification type algorithms such as K – Means
Clustering here.
K-MEANS CLUSTERING is a method of vector quantization, originally from signal
processing, that is popular for cluster analysis in data mining. k-means clustering
aims to partition n observations into k clusters in which each observation belongs
to the cluster with the nearest mean, serving as a prototype of the cluster.
We are calculating the mean of the values first of all, and followed by we need to
find the best accuracy (k) by providing some rough cluster values.
9. K – MEANS CLUSTERING Contd.
Once we received from Foursquare API neighbourhood values, we use k-means
clustering concept to form a cluster.
For making a cluster, we need to provide a cluster value which can be assumed.
We have created a variable for K-Means and assigning the same with location raw
data we have from various sources.
We are joining both the outcome from Foursquare API and K- Means.
Depends on the Cluster values, we could able to find the number of restaurants
available on the particular neighbourhood.
11. CONCLUSION
Depends on the values we have got from the clustering and segmentation
algorithms, we found clusters with values 1 and 2 [Central and Downtown
Toronto] would be suitable place to start a restaurant in.
If there is more vacant or non-resident areas in and around the venue, the number
may be small.
The investors should consider the competitive restaurants in the district.
Trying different kinds of foods which are not so familiar over there and advertising
their own products in an effective and consistent manner may increase the
possibility of getting good business in their neighbourhood.