Play around with DC real estate data
This data appeared in a job application question, where price should be predicted, with a test set was specified to be the rows of data that are a multiple of 5.
It looks like data that appeared on Kaggle:
https://www.kaggle.com/christophercorrea/dc-residential-properties/
See also:
https://www.kaggle.com/christophercorrea/preparing-the-d-c-real-property-dataset https://www.kaggle.com/together/eda-of-dc-residential-properties
And this example of using Keras in another domain (MPG prediction for cars):
https://www.tensorflow.org/tutorials/keras/basic_regression
Columns:
ID BATHRMNumber of Full Bathrooms HF_BATHRMNumber of Half Bathrooms (no bathtub or shower) HEATHeating ACCooling NUM_UNITSNumber of Units ROOMSNumber of Rooms BEDRMNumber of Bedrooms AYBThe earliest time the main portion of the building was built YR_RMDLYear structure was remodeled EYBThe year an improvement was built more recent than actual year built STORIESNumber of stories in primary dwelling SALEDATEDate of most recent sale PRICEPrice of most recent sale QUALIFIEDQualified SALE_NUMSale Number GBAGross building area in square feet BLDG_NUMBuilding Number on Property STYLEStyle STRUCTStructure GRADEGrade CNDTNCondition EXTWALLExtrerior wall ROOFRoof type INTWALLInterior wall KITCHENSNumber of kitchens FIREPLACESNumber of fireplaces USECODEProperty use code LANDAREALand area of property in square feet GIS_LAST_MOD_DTTMLast Modified Date SOURCERaw Data Source CMPLX_NUMComplex number LIVING_GBAGross building area in square feet FULLADDRESSFull Street Address CITYCity STATEState ZIPCODEZip Code NATIONALGRIDAddress location national grid coordinate spatial address LATITUDELatitude LONGITUDELongitude ASSESSMENT_NBHDNeighborhood ID ASSESSMENT_SUBNBHDSubneighborhood ID CENSUS_TRACTCensus tract CENSUS_BLOCKCensus block WARDWard (District is divided into eight wards, each with approximately 75,000 residents) SQUARESquare (from SSL) Xlongitude Ylatitude QUADRANTCity quadrant (NE,SE,SW,NW)