DQ report

[sinanshi]

• [x] table 1 - mostly done, do we still need SOFA socre?
• [x] missingness
• [x] file/site specified DQ report.
• [x] distribution
• [x] site specific.
• [x] data distribution comparison - with the old data average. In this case, should we clean the data first?
• [x] Create reports when there are new incoming files.

[sinanshi]

Table 1

Produce a biomedical style table 1 as part of our data quality report Where rows are key variables of interest which we summarise Two columns

one for existing data from that site
one for the new data

Plus a final 'comparison' column comparing standardised differences for each variable A bit like the propensity score match table to show that matching worked/didn't work

I have written a table generator here already

https://github.com/datascibc/dsbc/blob/master/R/table_one.R

And this file prepares and parses data

https://github.com/UCL-HIC/paper-brc/blob/master/analysis/table1.R

[docsteveharris]

next steps

• standardised mean differences as per matching package (later)
• visualise distributions for new vs existing
  • density plot with limits on x-axis
  • categorical ?? mosaic plots or simple bar graph (or just tables)

[sinanshi]

see #122