Microsoft Office Marketplace logo

Home | Download | Tour | Purchase | About

 

Natural language search and business intelligence

 

Here are a few business intelligence examples resulting from natural language searches. Let's take the following data :


Sample data.

 

If we type Sales ? in the keywords field, we see the following as output in the result list :


Magically computing sales by entering Sales ?

Entering Sales ? returns $690.00 as a result, which is effectively the sum of the Sales column. Please note the latter value does not appear anywhere in the spreadsheet. It is a calculated value. The search tool does the calculation.

 

If we type Sales by RegionID ? this time in the keywords field, we see the following as output in the result list :


Magically computing sales by entering Sales by RegionID ?

Entering Sales by RegionID ? asks the search tool to group sales by region and to calculate the sales for each region. There is more than one value being returned, that's why the search tool actually returns a table of values.

 

If your data has another measure of interest, such as Volume, you can combine them with "prix ht" by Volume and by Region ?


Combining measures by entering "prix ht" by volume and by region ?

 

If you are interested in another aggregation function such as min, max, average, count, you may explicitly ask so for instance with max Sales by RegionID ? :


Magically computing maximum of sales by entering max Sales by RegionID ?

 

For grouping in time, one of the 4 following keywords can be used : WEEK, MONTH, QUARTER, YEAR. As in sales by month ? :


Magically computing sales by month (note : there is no month column in the data)

Keyword list for natural language search :

  • aggregation function : SUM, MIN, MAX, AVERAGE, COUNT
  • grouping in time : WEEK, MONTH, QUARTER, YEAR
  • sorting : TOP, BOTTOM

 

...Continue to ...