Background

Our aim is to develop a BI assistant that can answer “BI questions”.

A BI question is defined as a free text question that has a single definite answer based on querying and aggregating data from a given dataset

For this assignment, we will be using the non-commercial dataset of IMDB (see here).

Here are some examples of BI questions on the IMDB dataset:

1. How many movies were released in the last year?
2. How many actors were born in x2001?
3. What is the language of the movie "Pulp Fiction"?
4. What is the series with the most episodes?
5. In how many horror movies did Sean Connery perform?

Part I

Propose a solution for building a BI assistant. Your proposal should include:

• Description of the high level architecture:
  • Frameworks
  • Models & Algorithms
  • Databases
• Description of any preprocessing stage: e.g. pipeline for data preparation, fine tuning..
• Description of the entire question answering flow

For the choices you’ve made, describe what alternatives you were considering if any and what are pros and cons

Part II

Implement the BI assistant solution you proposed in a simple command line application. your CLI app could be written in any language of your choice. it should be capable of answering the 5 questions example questions written above.

Submission instructions

• Please upload the code into a Github repo, along with a Markdown file describing your solution.