We know how to build graphs from keywords which is popular in text analysis. In this post, I discuss the use of IdeaMap using a graph of ideas but on a personal level and one that can be determined from a person’s activity monitoring in the form of email or written communication. 

Use Case:  Personal communications tend to be client-centric. A person sends out emails with responses that have his signature thoughts and ideas. On a day to day basis, a person in a particular vocation may have very specialized communication and deal with recurrent themes and ideas in a continuous basis. Such ideas can be captured in a pictorial representation with graphs. 

Giving the user this representation of his emphasis on a daily basis helps her be reminded of her impact areas and how she may need to tune her efforts in order to not miss on outliers and to make better channelization of her energy. This in a nutshell is the purpose of discovering ideas or themes and representing them in graphs on a daily basis as a form of personalization for the user. 

Today users can install a variety of machine data analysis toolsets like Splunk where we can feed all the documents received by the user to the inputs to Splunk and have charts drawn to highlight the most frequently used topics. However, the analysis suggested in this writeup goes far beyond that. To illustrate a few differences, we are interested in capturing a graph of topics for the day with edges and weights assigned to show interdependence and similarities. These are not the same as bubble charts in the visualization. Second, we are interested not just in the illustration but also the analysis to deeper than frequency counts of text chosen from input formatters. Specifically we refer to text analysis methods such as keyword detection and topic analysis. For example, we talk about normalizing all communications from the user to be represented as free text which we parse and analyze as a bag of words with a weighted matrix of keywords. Furthermore, we review the word embeddings and graph embeddings of the presented text to decipher the salient semantic content. We don't just stop there but also consider further semantic analysis as analogies, sequence completion and classification. Together,  these represent the best in the industry for deep learning of the communications from a user to present the topics of the day.  

The daily graphs can be a pictorial representation embedded in any of the documents or software used by the user for collaboration or referred to on a public website dedicated for this purpose through shortened hypertext or QR codes.  

Finally, the daily graphs can also be collected over time to be rolled up for monthly, yearly or other suitable periods. Since the graphs are accumulated over time, they present the opportunity to the user to drill down as appropriate.  

In conclusion, topic analysis in personalized communications and a visualization presented for summary are expected to delight the customer in reviewing with a glance the impact of the efforts spent over time. 

#codingexercise

#check divisibility by 9 for any permutation of a given number: http://ideone.com/tvGBxH