• pagetype : Comma seperated 'WEBPAGE_TYPE' values in 'lookup' data.
• metrictype : Requested statistics. Use 'fre' for time windowed frequencies, 'dur' for recency. Indicators must be comma seperated.
• timewindow : Comma seperated integers for time windowing.
• dateofreference : Date string to caculate how many days have passed since the last activity after this date reference.

• Small 'lookup' data BROADCAST joined.
• Memory partition by :"USER_ID","WEB_PAGEID" and date parts like year, month, day.
• Data cached before aggregation.
• Disk partition by "USER_ID".
• LOG4J for logging.
• unitttest for testing.
• python3

Example usage :

spark-submit PageViewStatisticsApplication.py --pagetype 'news, movies' --metrictype 'fre, dur' --timewindow '365, 730, 1460, 2920' --dateofreference '12/10/2019'