CSV Browser Demo

A web application that loads a CSV and enables searching through a large dataset to quickly locate data of interest. Build with React ⚛ & Python 🐍.

Goal

This project is to build an web application that:

• Accepts a CSV file of a predefined format (presumably log entries for tracking object changes)
• Allow users to query the dataset for the state of objects
  • At specific point in time
  • Base on specific attributes like object_type, object_id or both

Instructions

The easiest way to run this project is to start it locally using Docker.

Install and start Docker via command line.

$ docker-machine start
$ eval $(docker-machine env)

Take note of the IP address which docker machine runs in (you'll need it later).

$ docker-machine ip

Checkout this project and start the Docker containers.

$ cd demo-csv-browser
$ docker-compose build
$ docker-compose up -d

The project will run the following containers:

• frontend: Runs a webpack dev server with hot-reload enabled
• server: Runs a python flask server with hot-reload enabled
• database: Runs a PostgreSQL server. To access the database, use docker-compose exec database psql -U app

This will setup a development environment for the project locally. To view the project in the browser, you'll need the Docker IP and access it at port 8000, example http://[docker-machine-ip]:8000/

To stop the project, use

$ docker-compose down

Design

Several design considerations are taken into account:

• Users typically have no visibility to the time period of events within the log (CSV) file. Providing this information can quickly enable users to narrow the scope of search
• Users will begin the search process without knowledge of the specific time the event occurred. They usually have a broad range in mind (e.g. happened last quarter, last 1/2 year)
• Co-relating events is a very common use case in such operations, e.g. searching for event Invoice:231 that is part of an order event Order:45

Others

This project was developed under time constrain, and is not bug-free and requires much UI polish. Please use issues for suggestions and bug reports.