The repo contains sample data files (10K and 100K records) as well as the data file generator script (generate-records.py). The NG app simply displays data and allows the user to search/sort the dataset.
The dataset is in the following format: [ {"company":"COMPANY","email":"EMAIL","firstName":"FNAME","id":ID,"lastName":"LNAME","phone":"PHONE-NUMBER","state":"US_STATE"},... ]
The demo can be viewed here (10K, 100K, and 1M records): https://s3-us-west-2.amazonaws.com/angularjs-apps/large-dataset-demo-1M-records/index.html