Required work for a numeric field
- For any numeric field you use please make a worksheet containing
- A comment at the top describing what is in that worksheet.
- An entry on the About page describing the worksheet.
- Count:
- The number of non-empty cells (counta)
- The number of numeric cells (count)
- The number of blanks (countblank)
- The number of NAs (countif)
- Find
- Max value (max or quartile.inc(x,4))
- Q3 (quartile.inc(x,3))
- Median or Q2 (quartile.inc(x,2) or median)
- Q1 (quartile.inc(x,1))
- The minimum (quartile.inc(x,0) or min)
- Compute
- The average value (average)
- The standard deviation (stdev.P)
- Graph
- A histogram might be good.
- A modified histogram if your data is too spread out.
- If you suspect there may be outliers a box and whisker plot
- Any other graphs that you find useful.
-
- Make sure that you document the following in your methods document
- A general description of the field
- The five number summary.
- Possibly the other values as well.
- A chart depicting the values.
- A description of anything noteworthy.
- A description of any data cleaning required
- What was wrong
- How you decided to fix the problem. (High level how)
- Why you decided to fix the problem the way you did.
- The method you used to fix the problem. (Low level how)
- Any unresolved issues.
-
- This is a minimum, you may add more information you find useful.
What I Expect for Text Fields
For each text field used please create a worksheet
- Name the worksheet appropriately
- Add a comment at the top with a quick description of the worksheet
- Count
- The number of non-empty cells (counta)
- The number of numeric cells (count)
- The number of blank cells (countblank)
- The number of NAs (countif)
- For any text with a small number of unique instances
- Build a table of unique values
- Count the number of times each value occurs.
- Build a chart, perhaps a pie chart displaying the values.
- For text with a large number of unique instances
- Build a table of unique values
- Count the number of times each value occurs.
- Rank the data by occurrence.
- Build a chart showing the most frequently occurring values
- Perhaps count the number of unique, duplicated, ... values.
- Perhaps build a chart displaying the number of unique, duplicated, ... values.
- Provide any other analysis that seems important for your analysis.
- Add this information to your methods document.
- Discuss any interesting or important findings.