The code in this repo are primarily iPython notebooks for further filtering of CDMA PCMD files after being processed by the Network Coverage Visualization tool. While ultimately direct filtering of RAW PCMD records is desired so that the analyst can examine different relationships with data this is a pre-emptive step re-using the field mapping already performed by the NCV tool. However, the tool still delivers a significant size report in terms of metadata to be processed. Hence, the need for further manipulation to look for statistical outliers with respect to a report that describes approximately 132 concurrent data points per connection that was ultimately classified as drop. The goal is to extract the hidden relationships between the data points of each drop such that a pattern for correction in the most effective way can be perscribed. Author: S. Bryson, Ph.D. There are several notebooks in the directory each with the same theme of taking tabular data and sorting and formulating it such that it can be visualized or correlated among columns For CDMA the FullMarket notebook will process a complete market supplied data set from the NCV Call Drop study output. These files usually contain anywhere from 800K to 4.5M records (calls) from a major metropolitan market. There are other notebooks included that will process ill formed performance metric data from LTE KPIs. These data sets are usually ill formed in terms of sorting by hour, day, cell and technology. The python scripts re-factor the data to a format that is used to determine relationships and correlations.