Requirements And Usecases For The Data Platform
- The high-level goal of the project is to build an end-to-end service that can interface with ASIN or IDQ-level data and present the information to users through existing business reporting functionality in Winston.
Requirements
At a high level, the underlying database behind the data platform should fulfill the following criteria:
- High performance - we want to have the ability to query multiple vendor codes (or multiple attributes in general) and get a timely response. Currently, in Tableau, such a query results in tens of seconds of delays, and often crashes the Tableau dashboard.
- Effective sorting and flexibility - the data needs to be sortable efficiently, whether that is through a user-provided sorting pattern or a predefined default one. The user needs to be able to play with the data and ultimately be presented with only relevant information.
- Data needs to be downloadable as an Excel sheet/CSV for further analysis.
- It should be as convenient as possible to migrate and upload the data from S3 and Redshift.
- The provided solution should be as cost-effective as possible, while not sacrificing performance quality.
Use Cases
- At the most basic level, we want to be able to pass in a vendor code and time range, and show the relevant ASIN-level data (or the data relevant to the corresponding metric of the table)
- We should be able to ingest the following four datasets:
- ASIN Deep Dive
- IDQ Deep Dive
- Conversion dashboard
- Retail FastTrack
- For the ASIN-level data, the core parameters to consider are product family, product group, vendor code, MCID, and ASIN level.
- Users should be able to select which vendor code they want to be able to use as a parameter.
- We can expect a minimum of 500 users.
- We should be able to replicate the tabular view of the deep dive in Winston - Tableau’s charts are a secondary priority.
- The metric name also needs to be a separate parameter in the database, so we can easily ingest new metrics apart from the initial ones provided above. 202007090947
uid: 202007101008 tags: #amazon