1

Oct 14, 2013

Twitter Topic Explorer - Free tool


Jeff Clerk from Neoformix has come up with another beautiful tool called as Twitter Topic Explorer.

This tool does the text analytics on a person tweet. Based on word cluster algorithms, it classify the words used in tweets in different groups and display it beautifully.

You can explore the tool at http://tweettopicexplorer.neoformix.com

The tool retrieves the recent tweets of a particular twitter ID and extracts most common words from those tweets.

The size of the bubble is based on the word frequency and the words most often used together are grouped together with same color. This tool also works on the twitter lists.

We can use this tool in many ways:



  • As Jeff suggested, we can identify who to follow on Twitter. We can analyse who have similar  interests and are tweeting about the topics which are meaningful for you.
  • We can decipher the tweets of competitor and analyse what they are tweeting about
  • We can find out the topic trends based on our brand market influencers
  • We can correlate this data of our brand tweets with the other KPI’s like retweets, replies. This will help us in identifying the topic which is resonating with our customers and keeping them engage.


Below are some of the screenshots of the tool based on my twitter handle @gunjan_amit :



How you are planning to use this tool. Share your thoughts with us…


Sep 16, 2013

How to build Taxonomy for Social Media Analytics



According to survey done, billions of mentions on social media are on product and services of various brands. Through text analytics techniques, we can decipher the sentiments and dimensions of customer conversations around these products and services.

In my below posts, I have emphasized on the benefits of dimensions analysis on social media data:

  • Online and InStore Analytics using Social Media Data
  • Segmenting big unstructured data (Social Media) on different dimensions – Part 1
  • Segmenting big unstructured data (Social Media) on different dimensions – Part 2


Below process shows the steps involved in building the taxonomy:


  
Sample taxonomy will be like this:




We need to correlate this ontology with the sentiments and structured data like social platform author demographic information, Geo location, number of followers etc… to do deep dive social media analytics and to get actionable insights.

Do you face any challenge in building taxonomy? Please share with us and we will try to help you… 

Aug 13, 2013

Analyzing Big Data based on Segments

TV has become an integral part of our lives. Because of explosion of TV channels, it has become very competitive for Channel companies to attract attention of the viewers.
Channel companies are overloaded with various data points like viewership data, syndicated data, call center data, social media data, set up box data, broadband data etc…


Till we don’t integrate all these data sources, we cannot make a sense out of this BIG DATA.

The best way of integration is by SEGMENTS.


Below are the tables having audience segments of UK population:

Social Grade
Social Status
Occupation
A
upper middle class
higher managerial, administrative or professional
B
middle class
intermediate managerial, administrative or professional
C1
lower middle class
supervisory or clerical, junior managerial, administrative or professional
C2
skilled working class
skilled manual workers
D
working class
semi and unskilled manual workers
E
those at lowest level of subsistence
state pensioners or widows (no other earner), casual or lowest grade workers

Lifestyle/ description
TV channels watched
Other Media
Internet Use
Men 16-34
65% have no kids
60% are single
Majority work full time
80% have access to digital
Sky sports channel 4 bravo kerrang
25% use newspaper
Zoo nuts FHM
Cinema and radio
70% have broadband at home
Sport Banking News shopping
Men ABC1
12.8m in Britain
2/3rds are married
Scrubs, Simpsons, sport, news, documentaries
5 live, radio 4, tabloids newspapers less magazines
60% with broadband travel, sport, finance and news sites
Men C2DE
10.5m, constant media habits
ITV, five, Discovery, Bravo, The Sci-Fi Channel and Sky One
FiveLive and TalkSport Nuts and Zoo
44% have broadband
Men 55+
7.5 million, retired, more free time
ITV1, UKTV History, Discovery and National Geographic
Particularly radios 3, 4 and FiveLive. The Telegraph, The Mail and the Express
33% have broadband
Women 16-34
Broad set of media
Channel4, E4, ITV2, Living and MTV
Glamourg grazia, radio virgin and kiss
60% have access for shopping downloading
Women ABC1
Have children, more likely to be, single 30% of adults in UK
E4 UKTV Style Living, Channel4 ITV, the BBC
Cosmopolitan, Hello! Red
Over 50% have access
Women C2DE
11m broad range less than ₤17k 34% don’t work
ITV1/2 E4 and Living TV5
Sun Now, OK, Take a Break, The Mirror radio1,2, capital vigin
<35% have access
Women 55+
9m, free time
Day time TV bbc itv gmtv
½ read papers Daily Mail Mirror Classic fm
41% have access
Source : Wikipedia


Use cases:

  • ·         To provide customized ads based on the behavior of these segments and their viewership time
  • ·         To change the IVR options, by analyzing the trend of various issues on other touch points of customer
  • ·         To measure the impact of one segment over the other
  • ·         To train call center resource based on emerging trend
  • ·         To understand the impact of campaigns on various touch points
  • ·         To calculate ROI of social media efforts by correlating it with the viewership data
  • ·         To engage with the Right people at the Right time in the Right frame of mind.


Do you have any other use case to add? 
How you are integrating your customer touch points data together?



Jul 15, 2013

How to find influencers Klout Score


Using Klout, we can measure how the people who are active in social media influence each other online.

We can use Klout API to get the scores of people who are speaking about the brand on various social media platforms. These Klout scores can also be used as a benchmark to measure the engagement of brand’s owned media as compared to competitor’s social media handles.

Below is the process of generating the Klout scores using their API:

  1. Generate API key from URL http://developer.klout.com/apps/myapps
  2. Write API key in http://api.klout.com/v2/identity.json/twitter?screenName=barackobama&key=xxxxxxxxxxxxxxxxxxxxxx
  3. Write twitter handle name instead of “BarackObama”
  4. It will a generate ID
  5. Add that ID in http://api.klout.com/v2/user.json/2055/score?key=xxxxxxxxxxxxxxxxxx in place of “2055” and write API key in end
  6. It will give the Klout  score to you


 Image Credit : www.bitstrips.com

How you are leveraging Klout Scores in your social media strategy? Please share via comments....


Jul 7, 2013

Online and InStore Analytics using Social Media Data


This is in continuity of my previous posts “Classifying social media mentions on various dimensions - Part 1 and Part 2”.

Various retail brands have options of both online shopping and instore shopping.

For social media analytics monitoring, we have to make sure we do separate classification of dimensions for online and instore shopping experience of customers.

Below are some sample dimensions for Online shopping experience analytics:
  • ·         Price
  • ·         Delivery Time
  • ·         Quality of Products
  • ·         Pre-Ordering
  • ·         Customer Service
  • ·         Convenience
  • ·         Availability and Collection

Below are some sample dimensions for Instore shopping experience analytics:
  • ·         Price
  • ·         Quality of Products
  • ·         Staff
  • ·         Convenience
  • ·         Availability and Collection
  • ·         Billing Time



Jul 1, 2013

Social Network Analysis - NodeXL - Intro




NodeXL is a free and open source network analysis and visualization software package for Microsoft Excel (2007/2010).

You can download the NodeXL from www.nodexl.codeplex.com


Once you download the NodeXL, run it and you can see that as a separate tab in your Excel. Click on the tab and goto Import. You will see many options there. In this post, we will be focusing on “Twitter Search Network”.


Twitter Search Network will help in importing a Twitter network of people whose tweets contain a specified hashtag or word. For example, I have used here word “airtel”. When you are first time using the NodeXL, you have to get the app authorized with Twitter.



Once authorized, it will give you a PIN which needs to be added to NodeXL.




When the authorization is done, NodeXL will extract all the data in different sheets.


You can extract only one week old data of Twitter. You might need to connect to services like GNIP or Data Sift for achieve Twitter data. But still using NodeXL, you can keep appending the data to create a useful network data.

Click on “Refresh Graph” to create the network graph as shown in below pic.


In the above graph, you can see some 3-4 highly dense nodes. By clicking on Zoom and hovering over those nodes, you can find the twitter account for it.



Above graphs shows the dense nodes are of “airtelpresence” and “airtelug”. Both of these are brands own twitter handles.

Through this network analysis, you can find the influences who are twitting about the brand and have a high reach.

In the upcoming posts, I will be covering the other features of NodeXL.

Through comments, let me know your query on NodeXL. 

Jun 19, 2013

Linkedin Network Analysis

If you are passionate about Social Network analysis like me, then this post is for you.

Try out creating social networks for your linkedin connections using http://inmaps.linkedinlabs.com/

Below is the social network for my profile


I will be doing the deep dive analysis on this in later posts. Till then enjoy this beautiful free Linkedin network analysis!



May 24, 2013

SPSS Text Analytics for Surveys – Tips and Tricks 6

Here is one more post in this series of best practices/tips on SPSS Text Analytics.

There is a possibility that the tool has missed any important sentiment which is important for you. But directly from the response, you can select any particular word and then add the sentiment associated with it.

Select the word from response  -> Right Click -> Add to Type -> Select relevant sentiment



Please share your best practices/tips of Text Analysis !!! 


May 5, 2013

RACI Matrix for Web Analytics Project



If you are doing a web analytics project in collaboration with a tool vendor, then this post is for you.
To make the project successful it is always better to remove all ambiguities and confusions before the project starts.

Here I have created a RACI matrix to cover all the phases of Web Analytics (Discovery, Design, Deploy, UAT and Reporting).

RACI stands for Responsible, Accountable, Consulted, and Informed

For more details, please refer http://en.wikipedia.org/wiki/RACI_matrix

The first step is to create a project plan and this RACI matrix. Get it signed off from vendor and your client before you starts the project.

Responsible
Accountable
Consulted
Informed
Business Requirement Document
Vendor
Vendor
Client
Client
Solution Design  Document
Vendor
Vendor
Your Company
Client
Technical Design Document
Vendor
Vendor
Your Company
Client
Test Cases
Your Company
Your Company
Vendor
Client
Audit Report
Your Company
Your Company
Vendor
Client
Administrative Work of tool
Your Company
Your Company
Client
Client
Deployment of Tags
Your Company
Your Company
Client
Client
Training
Your Company
Your Company
-
Client
Data Feeds (if Required)
Vendor
Vendor
Client
Client
Reporting
Your Company
Your Company
Client
Client

Please comment if I am missing any point here.


Mar 16, 2013

Social Media Segments for Retail Industry


In today’s time, Social Media is a best place for Retail industry to explore and act. With the growth of smartphones, customers are expressing more on social media. Customers are writing about the product, customer service, bargains/offers/coupons etc… There is rich information available on social media platforms which Retail industry can leverage and integrate that with their other data sources.

The challenge is to find the demographic information of the customer for creating personas as many mentions are by “Anonymous” visitors or because of privacy settings of social networks.

But we can still do the persona analysis on social media data by using Text Analytics techniques.

Some of the segments which can be created using social media data are:

  • Price Conscious Shoppers
  • Loyal Shoppers
  • Grab and Go Shoppers


Using the Text Analytics power, we need to first create a bag of terms which will help in identifying the customers of each segment.

Some keywords which can be used for each segment are:

  • Price Conscious Shoppers - Offer, Coupon, Sale, Discount
  • Loyal Shoppers – Love Store, favorite Store, Best Store


Sample Mentions for these segments are:

  • Price Conscious Shoppers – “Those bargain baskets in the front of the xxxxxx store always get me!”
  • Loyal Shoppers = “I love shopping in xxxxx store”
  • Grab and Go Shoppers – “Will pick up bread from xxxxx store on my way to home”


When you are done with creating customer segments on your social media mentions, integrate it with the CRM segments and sales data. Correlate the trend and analyze. 

  • Is there any impact of tonality of any particular persona on sales data? 
  • Do targeted segment is not responding to your campaign? 
  • Why any particular segment is behaving in a strange manner? 
  • Which segment you are going to target in next campaign?
  • Is there any correlation of your CRM segments with Social Media Segments?




The business benefit is that it will help the brand to make campaigns smarter and calculating the ROI of social media marketing strategies.


Related Post

Related Posts Plugin for WordPress, Blogger...