Ravindra BagaleCourses & study guides

28. End-to-End Project: Blinkit Quick-Commerce Dashboard

28.3 Step 2 – Get Data and Transform in Power Query

  1. Create the parameters FolderPath and StartDate (Module 10). Get data › Folder (or SharePoint folder) › FolderPath › filter the files › Combine (Module 9). Name the query Blinkit_Orders.
  2. Get data › Text/CSV › AmazonNow_Orders.csv › Transform Data. Rename Order_No → Order ID, Order_Value → Amount, and so on, so both queries have identical column names.
  3. In both queries:
    • rename raw columns to friendly names (OrderDateTime → Order DateTime, DeliveryTimeMins → Delivery Time Mins …);
    • set data types: Order DateTime and Delivered DateTime as Date/Time, Quantity and Delivery Time Mins as Whole number, Amount/Discount/Delivery Fee as Fixed decimal;
    • add Order Date (DateTime.Date), Order Hour (Time.Hour) and Delivered Date columns; then remove Order DateTime if it is not needed (reduces cardinality);
    • Trim and Clean the ID columns; standardise (एकसमान करणे) city spellings (Chh. Sambhajinagar, Aurangabad → Sambhaji Nagar) with the CityMap mapping-table merge (Module 7.9);
    • flag invalid values (negative quantity, delivery time 0 or above 180 minutes) with a Data Quality conditional column (Module 7.16);
    • add a Platform column if the file does not already have one.
  4. Append Queries as New (Blinkit_Orders + AmazonNow_Orders) → Orders; untick Enable load on the two source queries.
  5. Load Product, Customer, DarkStore, DeliveryPartner from Excel; check headers, types, and Remove Duplicates on the key columns.
  6. Use Column quality/distribution (profiling on the entire data set) to check errors and empty values. Expect blank Delivery Time Mins only for cancelled orders.
  7. Data-quality check: a Left Anti merge from Orders to DarkStore finds Store IDs missing from the store master.
  8. Close & Apply.

Ravindra Bagale's Tip

Mitrano, khup students clean the data, load it and never check the result against the source. After this step, compare row counts and total Amount for one city and one day with the original export. Fix differences now, before the model and visuals depend on them. Dhyan rakho!