Spatial Data Science
How to Choose the Ideal Site for Designing Your Restaurant Using Data Science
Designing any product requires a lot of analysis and research. Restaurants are no different.
Designing any product requires a lot of analysis and research. It is also true for designing any building. Before we begin to design any building, we collect information about the location where we are designing, we check the budget, we do research about the background of the target audience and their likes and dislikes.
In designing any space you need to follow some preliminary steps, like budget analysis, client requirements, form & concept development, site analysis, and zoning.
Site Analysis is one of the major steps in the pre-designing processes. It involves analysis and evaluating topography, watercourses, trees, manmade features, site boundaries, habitats, wind circulations, connectivity to main roads, streets and pathways, distance to closest facilities or amenities, climatic conditions, and weather patterns.
In my Architectural Design course, we were given the assignment to design a restaurant in our home town, with very minimum and basic specifications. Having complete freedom to choose the site, concept, form, building materials, and cuisine offered, it required a lot of thinking and creativity.
But it is very difficult to locate the ideal location in such a big city as it also requires a huge amount of research to locate a site because the restaurant should also be successful in business after it starts to function. So instead of choosing my favorite or city’s most popular restaurant’s or cafe’s location as my site, I thought of implementing Data Science for analyzing the data of the restaurants in Pune (my home city, in the western Indian state of Maharashtra) and then using Machine Learning to locate the ideal site.
For Data Science you need a large amount of data for the results to be precise, so I collected data about restaurants in Pune through Zomato API’s, population data, pollution data, and keeping in mind the current situation, I even used the geospatial data of COVID-19 cases, from various available sources.
But the COVID-19 data was of not that use as the number of cases and containment zones change every day. So whatever area the Restaurant is in, precautions should be taken.
For the analysis, I used the following factors to locate the ideal location in the city:
- Ratings
- Votes
- Geolocation of the Restaurant
- Locality / Neighbourhood
- Price Per Sq. Feet of the Commercial plot of each Neighbourhood
- Cost for two people
- Type of the Restaurant
I observed and analyzed each and every factor for location analysis and prediction, and only considered the factors that were important for me. There can be many more factors but due to data availability constraints, I used the above-mentioned factors only.
Locating the ideal site is important but it is not the only factor to be considered in the planning phase, you should also decide what you are going to serve, price of the land or the rent and what should be the ideal cost. These factors should also be analyzed wisely as only locating the site will not get you customers and rating it is also important where you serve what and for what price.
For example, if the ideal location based on your budget is the “Business and Office” area then Casual Dining would be an appropriate Type and the cost for two people can be more so you can invest more things like formal furniture and menu, if you are planning for a restaurant in an area which has many schools, colleges, and coaching classes then Cafe might be the best alternative as students prefer cheap food and a good place for hanging out. This is just a rough idea or an example, before actual analysis.
Popularity and Rating are two different factors for judging a Restaurant, popularity can be seen through the number of votes. When it comes to popular localities Kothrud is the most popular locality for cafes, restaurants, and eateries as it is a densely populated area with several colleges and schools. Kothrud is followed by Viman Nagar and then Hinjawadi. Hinjawadi is a corporate location consisting of a high number of offices.
Locality or neighborhood or suburb are all broader terms. This is the list of top 10 specific locations that serve good food as they have a rating of above 4. Rating symbolizes customer satisfaction whereas the number of votes tells us the average footfall of that restaurant.
This graph tells the average rating of each locality and is ranked accordingly. Kothrud is a popular eatery hub, but in terms of great food and good service, Baner occupies the top position. This can be a crucial factor while deciding the location for a restaurant.
Not every type of restaurant works everywhere. Popularity and Rating are major factors for deciding the ideal location but the type of restaurant also plays a crucial role in success. Kothrud might be the most popular locality in Pune but not necessarily every restaurant type will work there. For example, Casual Dining has the most number of votes in Viman Nagar, Dessert Parlour, or Bar has popularity in Baner. The popularity of a type in a specific location depends on the type of audience living there. Baner and Aundh are popular hangout places for a younger generation so Microbrewery, Pub, and Dessert Parlour gain more audience here.
The cost of the food never decides the popularity or rating of a restaurant. But this graph shows an ideal range price depending on the rating. For the rating to be 4.0 < the ideal cost for two people should range from ₹600 — ₹1200. No restaurant in Pune was ever rated 5.0.
When locating an ideal restaurant site is done the next time aspect for consideration is the budget. The price of commercial lands is not even close to similar. Wakad has the costliest commercial land, followed by Balewadi. The price of the land does not play any role in deciding the success of the restaurant but budget planning is also important while choosing the location.
This gives a general idea of which type of restaurants are popular in Pune. Quick Bites and Casual Dining are popular in Pune, the city being an educational hub as well as the IT hub.
I trained this data in one of the Machine Learning algorithms and deployed the web application. Users can choose the input values as per their choice and the ideal locality based on the input values is predicted. (I have deployed the web application using shiny.io)
