analysis/cleansing: initial structure created

Author: sanand0

analysis/cleansing.md

@@ -0,0 +1,23 @@
+Data cleansing
+==============
+
+Often, the data you have will have errors, and you'll need to clean it up.
+Missing data, inconsistent labels, duplicates and misspellings are common
+examples of errors. The [Wikipedia article on data
+cleansing](http://en.wikipedia.org/wiki/Data_cleansing) talks more about this.
+
+The typical cycle of data cleansing is:
+
+1. Find errors
+2. Correct the errors
+3. Repeat 1 & 2 as often required
+4. Automate the correction process for new data feeds
+
+Finding errors
+--------------
+
+Correcting errors
+-----------------
+
+Automating correction
+---------------------