Survey banner
The Dataiku Community is moving to a new home! We are temporary in read only mode: LEARN MORE

CHALLENGE #2: Bangalore Online Food

InesB
Dataiker
CHALLENGE #2: Bangalore Online Food

You are the successful owner of a well-known fast-food chain in India, and you want to open a new restaurant in the dynamic city of Bangalore. 

Given that most of your customers order online and one of your chain's key strengths is its unbeatable delivery times, it’s crucial to select the optimal location for your new venture.

To make an informed decision, you’ll need to identify the areas in Bangalore where the current service offerings are the most disappointing, as evidenced by the highest proportion of negative feedback. This strategy will help you attract the most customers by providing superior service in these underserved areas.

You have two datasets to work with:

  1. Online Food Orders and Feedback: This dataset includes information about online food orders in Bangalore, including customer feedback and geographical coordinates of the orders : https://www.kaggle.com/code/tayyabli/online-food-dataset/input
  2. Bangalore Areas Geographical Data: This dataset contains geographical coordinates of different areas in Bangalore : https://www.kaggle.com/datasets/hegdetapan/bangaloreareaspincodewithlatitudelongitude

Your challenge is to use geographical tools to visualize a map that highlights the neighborhoods with the highest proportion of bad reviews.

You can accomplish this by following these steps: 

  1. In both datasets, create a column with Geopoint (using longitude and latitude columns) 
  2. In the Area dataset, use the Generate a circle tool to draw a circle around the geopoint you just created (with a radius of 1km)
  3. Match each order location with the corresponding Bangalore area by performing a Geojoin.
    1.  Use a Left join mode where the client’s geopoint is contained within the circle area.
    2. For the output dataset, select the following columns: Feedback, Area, and Circle Area.
  4. Calculate for each neighborhood the part of positive feedback
    1. Use a prepare recipe to compute a count of 1 for each row where the feedback is positive (Count number of occurrences processor)
    2. Use the Group Recipe to aggregate the number of positive feedback by neighborhood.
    3. Calculate the proportion of of positive feedback for each neighborhood with the Create column with formula tool
  5. In the output dataset, create a Geometry map to visualize the neighborhoods and add your part of Positive Feedback as a feature. You can then play with settings and filters of the chart to identify the few neighborhoods with the worst feedback. 

Ensure that all new datasets created during this process are saved in CSV format for compatibility with geographic tools.

By completing this challenge, you will be able to pinpoint the neighborhoods where customers will be most pleased to welcome a high-quality fast-food option.

InesB_0-1717778187924.png

Once you find the solutions please feel free to export your project and upload it here so we can all benefit from each other's efforts ! 

 

0 Replies

Labels

?
Labels (2)
A banner prompting to get Dataiku