Spatial data analysis

i have data with longitudes and latitudes. In that about 30% data is having accuracy more than 1 km i.e. 1000 meters. So what can i do like whether i have to delete those records or any other way.
i have to build cluster model on top of it

Hi ,

Since you can not delete data , We should check if the outlier is due to wrong recording of data when data collection had happend or the sample we took not having data points close to these outliers. If the outliers are due to wrong recording then we should correct and proceed and in case if the outlier is because of the wrong sample (I mean only one or two recording close to this data point exist in sample but population has many such observations and unfortunately we missed them in sample. So in this case we should add more data points which exist close to outliers).

© Copyright 2013-2019 Analytics Vidhya