Dynamic data transform in data visualization

Question

I have a question about data visualization in java. I work for marketing team and I need to build report for them.

My model:

class AdReportEntity {
    String date;
    String clientName;
    AdInfoEntity adInfo;
    int cost;
    int impression;
    int click;
    ...

}

AdInfoEntity:

class AdInfoEntity{

String campaignName;
String adgroupName;
String adname;
...

}

Client side have 2 main component:

1. bar chart show daily cost (sum of all client).
2. the table show the detail:
   date | clientName | campaignName |adgroupName| adName| cost| impression| click

The problem:
The requirement is more dynamic chart and table. Some time user want to display the chart by Client, or by AdName, or by campaignName, or maybe just group by first 3 char of campaignName, etc.
How to implement the server for more dynamic?

My solution is to SUM base on the param user request,Ex:

 String groupColumn=request.getParameter("column");
    switch(groupColumn){
    case "clientName":
        doSumByClientName();
    case "adName":
        doSumByAdName();
    case "spec1":
        doSumBySubStringAdName():
        ....
    }

I have the same problem for sorting and filtering.
I'm also tried column base for my entity but it's hard to deal with depend properties like ctr=click/impression.

Answer 1

You might take a look at Tablesaw, (https://github.com/lwhite1/tablesaw), Not sure it does everything you want but it certainly does sorting and filtering in Java. Here's some code.

Table tornados = Table.createFromCsv("data/tornadoes_1950-2014.csv");
tornadoes.setName("tornadoes");

tornados.structure().print();
output:
    Structure of tornadoes_1950-2014.csv
    Index Column Name Column Type 
    0     Date        LOCAL_DATE  
    1     Time        LOCAL_TIME  
    2     State       CATEGORY    
    3     Scale       SHORT_INT   
    4     Injuries    SHORT_INT   
    5     Fatalities  SHORT_INT   
    6     Start Lat   FLOAT       
    7     Start Lon   FLOAT       
    8     Length      FLOAT       
    9     Width       FLOAT    

tornados.shape()
output:
    59945 rows X 10 cols

// gets the month from every date, and counts by month
CategoryColumn month = tornadoes.dateColumn("Date").month();
month.summary();
output:

    Column: Date month
    Category  Count 
    JANUARY   1417  
    MARCH     3935  
    JULY      6305  
    MAY       13072 
    SEPTEMBER 3142  
    OCTOBER   2332  
    DECEMBER  1355  
    NOVEMBER  2364  
    FEBRUARY  1672  
    APRIL     8533  
    AUGUST    4066  
    JUNE      11633 


Table injuriesByScale = tornados.median("Injuries").by("Scale");
output:
    Median injuries by Tornado Scale
    Scale Median [Injuries] 
    -9    0.0               
    0     0.0               
    1     0.0               
    2     0.0               
    3     2.0               
    4     14.0              
    5     141.0             

// query:
Table fatal =  
     tornados.selectWhere(column("Fatalities").isGreaterThan(0));
// sort
Table sorted = tornados.sortOn("State");

It does lots of other stuff, too, including some stats, simple plots, etc.

Full disclosure, I'm the main author of Tablesaw, and it's still a bit rough around the edges.