Performance Task: Investigate --- November 2013


I have (I think) Performance Task: Investigate down to a science.
1.   Insist on some kind of transactions raw data --- something with a time stamp for each record. Why? Because transactions have fields and records that fit nicely into a pivot table.  Pivot tables are really good for creating questions and defining a focus that is scalable.
2.   Insist on CSV format.  Why? Because that format is easily readable by programming languages and by Excel.
3.   Do a pivot table in Excel (quicker than programming) --- cross tabulate in such a way (if possible) that all cells have meaningful numeric values and the summation of each row has a meaning and the summation of each column has a meaning.
4.   Add three to five questions and define the focus.
5.   Do in a programming language if time permits.
These steps are working in my classroom (I think) as of this writing. I think I first heard of pivot tables from Roy and Jill.
I will have to try some of the TED talk stuff mentioned above --- I watched it and it was certainly very interesting.  There is a pivot table chart feature in Excel that I have not had much time to explore.


Here is a somewhat abstract comparison of MapReduce and pivot tables:

MapReduce is a Google technique for processing big data.  Both MapReduce and pivot tables map and reduce data.  What are the differences?
MapReduce works using a programming language and pivot tables work in Excel.
MapReduce works with messy data and pivot tables do not.
MapReduce can parallel process data too big for one machine and pivot tables cannot.
Some of the database operations may vary slightly.
There are a lot of similarities.  I think doing pivot tables to eventually introduce students to MapReduce is a workable plan. One neat thing about pivot tables --- the user can double click a cell and see all the data that was reduced for that cell.  When you see the data that was reduced for one cell, the whole operation is illuminated.