pivot-chart-demo's Introduction

Pivot Chart Demo

This is a project that uses ElasticSearch, Flask, and the d3 library XCharts to demonstrate pivot charts and graphs based on aggregation queries to ElasticSearch. The front-end uses BackboneJS and Twitter Bootstrap. The backend runs on Tornado web server and Flask.

Requirements

You need to have the latest ElasticSearch (tested on 2.1.1) installed to run this project.

Setup

Check out a fresh copy and enter the directory:

$ git clone https://github.com/chaimpeck/pivot-chart-demo.git
$ cd pivot-chart-demo

Install the Python dependencies:

$ pip install -r requirements.txt

You need to set up the schema in ElasticSearch (which should already be running):

$ curl -XPUT "http://localhost:9200/pivot-demo/" -d'
{
   "mappings": {
      "entry": {
         "properties": {
            "gender": {
               "type": "string"
            },
            "age": {
               "type": "integer"
            },
            "household_income": {
                "type": "integer"
            },
            "favorite_color": {
                "type": "string",
                "index": "not_analyzed"
            },
            "QS6_6": {
                "type": "string",
                "index": "not_analyzed"
            },
            "Q30b": {
                "type": "string",
                "index": "not_analyzed"
            },
            "Q31_1": {
                "type": "string",
                "index": "not_analyzed"
            }
         }
      }
   }
}'

Generate 100k sample entries:

python generate-random-data.py -n 100000  > sample-data.json

Load the sample data into ElasticSearch:

python load-es.py sample-data.json

Run the demo:

python pivot_demo_service.py -p 8080

You should now be able to view the demo in a web browser: http://localhost:8080

If you are accessing it remotely, you will need to supply the username/password: pivot/demo

The Front-end

The front-end consists of three views:

Pivot Table: Demonstrates pivot tables that dynamically generated based on a query to elastic search
Filters: Allows user interaction to modify the filters by adjusting sliders and checkboxes. The graph changes with the user's actions
Search: Send raw queries to elasticsearch - mostly useful for debugging.