1

Dec 26, 2012

Schema in Adobe Insight


As per Wikipedia,

A database schema of a database system is its structure described in a formal language supported by the database management system (DBMS) and refers to the organization of data to create a blueprint of how a database will be constructed (divided into database tables). The formal definition of database schema is a set of formulas (sentences) called integrity constraints imposed on a database. These integrity constraints ensure compatibility between parts of the schema.

If you are starting new in Adobe Insight, then the first thing is to understand Schema.

You can view the Schema Diagram in Adobe Insight by

  • Right Click on Workspace
  • Click “Admin” -> Schema Diagram





Some Basic Adobe Insight Definitions:

User: A person who navigates to your site or uses your business
Interaction: How a person interact with your business in one sitting
Event: Any event that occurs during a person interaction with your business


Dec 16, 2012

How to Add Annotations in Adobe Insight?


In Adobe Insight, you can add easily annotations in any chart or table.

Through Annotation feature, you can add any text or image which can help you in noting down any particular trend, event, observation or any other noteworthy item.

How to add annotations in Adobe Insight:

  • Right click on your Workspace.
  • Select “Add Callout” -> Annotation -> Image/Text as shown in below pic.


Jul 29, 2012

Tweets per Day Analysis



This is another awesome tool “Twitter Venn” from Jeff Clerk.  

You can enter the brand and competitor names separated by commas and search. After the data is retrieved a Venn diagram is shown which illustrates the rate of tweets containing the search terms in the various combinations.

As per Jeff Clerk, this tool supports investigation into the relationship between how words are used within the messages of all the people using Twitter.

It also helps in visualize the overlap between various sets of topics.




When you click on the bubbles, it will show your related keywords cloud on bottom left hand side. Whereas on bottom right hand side you can see the average tweets per day.

Also on clicking the bubble, you can see the original tweets. Like when I clicked on the Brown color bubbles which represent the common tweets of Wipro and Cognizant, below tweet appeared:




You can analyze that the tweets having topics Cognizant and Infosys are more as compared to other combinations. This will help you analyze the other topics associated with your business lines. It will help you in making your marketing strategy more effective.


Do you have any other similar tool to share? Please tell us about your tool via comments.

Jul 26, 2012

Twitter Words Association Analysis



Recently I came across Twitter Spectrum tool from Jeff Clerk. This tool is modified version of News Spectrum tool.

Here you can enter two topics and then analyse the associated words based on twitter data. Blue and Red color represents the associated words of those two topics whereas Purple represents the common words.

You can click on any word to see the related tweets. The visualization is really awesome and you can easily analyze the data.

For example, I have taken “icici” and “hdfc” as two topics. Below is the twitter spectrum based on these two topics:

If you analyze the associated words, you will find that “Security”, “Insurance” words are associated more with ICICI bank whereas “digital”is associated with HDFC bank.

But words like “online banking”, “mobile banking” are still missing. This type of visualization will help the brand to plan their startegy of type of tweets to be posted on twitter.




Do you have any other similar tool to share? Please tell us about your tool via comments.

Jul 22, 2012

To find frequency of the words using RapidMiner



In my previous post, I wrote on How to read and write data in RapidMiner. In this post, I am covering How to count the words frequency in text using RapidMiner. The model contains following operators:

  1. Read Excel
  2. Nominal to Text
  3. Process Documents
  4. Tokenize
  5. Transform cases
  6. Filter Stopwords


RapidMiner model is shown below:



In Process documents operator, add 3 operators as shown below:



Tokenize operator splits the text of a document into a sequence of tokens.

Transform cases operator transform the words cases in desired format.

Fiter Stopwords operator removes English stopwords from a document like and, or, not, is, an etc…

Output :


If you are looking for XML of this word frequency model using RapidMiner, leave your email ID in comment box.


,

Jul 8, 2012

How to Read, Write data and Transform Cases in RAPIDMINER



I recently started exploring RapidMiner to do sentiment analysis and text classification of social media data. So I am going to post some tutorials on RapidMiner based on what I have learned so far on this tool.
In this post, I am writing on very basic thing – How to read, write data and transform cases in RapidMiner.

RapidMiner is a free tool and can be downloaded from  www.rapid-i.com .  Make sure you have Text Analytics plugin of RapidMiner installed.

Below is the model, I have built in RapidMiner to read and write text.

It includes 5 operators
  • Read Excel
  • Nominal to Text
  • Process Documents
  • Transform Cases
  • Write Excel



We start with Read Excel operator. “Read Excel” operator loads data from MS Excel spreadsheets. This operator is able to reads data from Excel 95, 97, 2000, XP, 2003 and 2007.


Select the excel file from your system which you want to load.





Excel file which I have uploaded using Read Operator is shown below.



Connect it with the “Nominal to Text” operator.  This operator replaces all nominal attributes by corresponding string attributes.

Then connect it further with “Process Documents to Data”.

At the end connect “Write Excel” operator. Select the excel file where you want to write the output. This operator can be used to write data into Microsoft Excel spreadsheets. This operator creates Excel files readable by Excel 95, 97, 2000, XP, 2003 and newer.



In the process documents, I have added operator “Transform Cases” as shown in below pic.
“Transforms case” operator transforms all characters in a document to either lower case or upper case, respectively.




Select from the drop down whether you want to transform all text to upper or lower case.




Run the Model.


You can see that the text have been transformed to lower case and will be written in output excel file.




If you are looking for XML of this model, please leave your email ID in comment box and I will share the XML with you.

Jun 3, 2012

SPSS Text Analytics for Surveys – Tips and Tricks 5


In our earlier posts, I wrote on how to add any particular key entity into the category.

The other method is that you can add directly from the response pane. When you have done almost all the categorizations and left with few. Browse through those uncategorised responses and read them.

After reading if you find that any key entity concept should come in particular category then select that entity and right click -> Add to category.



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

May 29, 2012

SPSS Text Analytics for Surveys – Tips and Tricks 4


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

If you want to mark any response with flag for future response, then right click on that response -> Important Flag.


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

May 23, 2012

Exclude Internal Traffic in Google Analytics Reports


When we are using Google Analytics to track clickstream data, we should segment the internal traffic and ouside traffic for our analysis. We all mainly use IP addresses to exclude the internal traffic. But it sometimes become difficult when we have the offices in different geo’s.

Here I am using “Service provider” report data to exclude all the traffic from internal networks.
Goto Audience -> Technology ->Network

Filter out all the internal networks by your company name.  You can see all your internal networks there.


You can either create an advanced segment or create a new profile excluding the internal network traffic.

Advanced Segment

  • Goto Advanced Segments.
  • Select “Exclude” from first drop down
  • Select “Service Provider” from second drop down
  • Then select “containing” from third drop down as shown in below pic.
  • Write your company name in the text box





Create a new profile

  • Click “+New Profile”.
  • Give name to your new profile.
  • Create New Filer.
  • Give name to your new filter.
  • Select “Customer Filter” and then select “Exclude” option.
  • From “Filter Field” drop down select “Visitor ISP Organization”.
  • Write your company name in “Filter Pattern” and save it.








Do you know any other way to exclude visitors from internal networks in Google Analytics reports?

Please share via comments....

May 13, 2012

SPSS Text Analytics for Surveys – Tips and Tricks 3


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

If you want to copy all the responses into excel file. Then right click on response pane -> Select All -> Copy

It will copy all the responses with categories defined for each response.




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

May 11, 2012

SPSS Text Analytics for Surveys – Tips and Tricks 2


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

If you find any particular survey text which is not coming under relevent category, you can force that response into required category.

For that, right click on that response -> Force Response Into -> Select desired category


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


What will be the Best Time to Tweet?


As per the bit.ly study, posting between 1 – 3 pm EST Monday through Thursday is best for achieving a high click count. But tweeting after 8:00 pm and as well as posting links after 3:00 pm on Fridays should be avoided.

Image Courtesy: http://bit.ly/JcirJE

But is it really any fix time to get the most traction on Twitter? 

NO

You have to tweet based on your market segment. Who are your customers? In which time zone are you customer in?

One should tweet based on time zone in which customers are there. Also it depends on what is the persona of those customers are?

For example, one should tweet based on their target customers personas which can be housewives, teenagers, business executives, retired people etc. Analyze the habits of your target customers. Then map it with their time zones. That will give you the best time to tweet.

But what would be the case when you are targeting the worldwide audience. I would suggest analyzing yourself the best time to tweet.

Post the three different tweets on different time zones for one month. Then analyse the interactions/clicks on daily basis as shown below:



This will give you the insights that when you are getting the most clicks on twitter.

The next step is to make the similar chart for each category of your tweets. If you are from B2B industry then segment you tweet into different categories like:

  • Company Specific News
  • Company Blog Posts
  • Training/Events/Conferences
  • Domain Specific Articles


If you are from B2C industry then segment you tweets in categories like:

  • New Products
  • Offers/Deals/Discounts
  • Brand News
  • Industry Trends



This will help you in deciding what type of content is getting clicked at what time. Based on the above analyses you can decide content and posting time of your tweet.

Right content to right people at right time at right platform is KEY TO SUCCESS.

There are many tools available to decide on best time to get most interactions on tweeter. Some are listed below:













Timely






Apart from this, there are many tools available to distribute your tweets over time like Media Funnel


If you want me to send the "Best Time To Tweet" dashboard, please leave you email ID in comment box. I will email it over to you.

Let us know via comments what time works for you depending on your market segment?


May 9, 2012

SPSS Text Analytics for Surveys – Tips and Tricks 1


Here is one more post on "How to rename the categories" in the series of best practices/tips on SPSS Text Analytics.

After creating the categories, if you want to rename any category then right click on that and click on “Rename Category”.






Write a new name for the Category.





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


May 8, 2012

SPSS Text Analytics for Surveys – Best Practices : 5


In my previous posts, I have written on How to start using SPSS Text Analytics tool and How to do categorization.

Here I am sharing some more best practices/tips on SPSS Text  Analytics tool.

If we wanted to export our categories then go to -> Menu -> Categories -> Manage Categories -> Export Categories



Define a file name and location to which the category data will be exported.


Choose the format for the category data that will be exported. You can choose between Flat/Compact list format or Indented format.


Review your output.

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

SPSS Text Analytics for Surveys – Best Practices 4


In my previous posts, I have written on How to start using SPSS Text Analytics tool and How to do categorization.

Here I am sharing some more best practices/tips on SPSS Text  Analytics tool.

If you want to add the key concept entity into particular type or pattern then just right click on it -> select “Add to type” -> Select the type (in below pic I have selected the Location type)


May 6, 2012

SPSS Text Analytics for Surveys – Best Practices 3

In my previous posts, I have written on How to start using SPSS Text Analytics tool and How to do categorization.

Here I am sharing some more best practices / tips on SPSS Text  Analytics tool.

If you wanted to copy the entire list of key concepts from the tool, then select all the entities. Then right click on it and click copy as shown in below pic.


This list you can paste in excel for your reference and analysis.

May 5, 2012

SPSS Text Analytics for Surveys – Best Practices - 2



In my previous posts, I have written on How to start using SPSS Text Analytics tool and How to do categorization.

Here I am sharing some more best practices/tips on SPSS Text  Analytics tool.

As we know the tool extract the key concepts in starting of the analsys. But some of those might not be relevent to us. But we can easily exclude these extractions from our list.

To do that just right click on that particular entitiy and click on “Exclude from Extraction” as shown in below pic.


Please feel free to share best practices or tips on Text Analytics!!!

SPSS Text Analytics - Best Practices

In my previous posts, I have written on How to start using SPSS Text Analytics tool and how to do categorization.

Here I am sharing some more tips on SPSS Text  Analytics tool. 

If you wanted to see how the tool has done basic categorization of key concepts entities in corrleation with the sentiments, then
  • Select the “Color Type” from the first drop down
  • Select “Type Pattern” from the second drop down

Please refer the below pic :





Through this you can quickly analyse which one has the highest frequency in your survey and focus on that first.

Please feel free to share best practices or tips on Text Analytics!!!

Apr 30, 2012

SPSS Text Analytics for Surveys – Tips and Tricks



Here I am sharing some tips on categorizations which will give useful and meaningful insights to your data.

You might wanted to know that how many said positive or negative about the food category. 

To create category for this go to Advanced catgorization. Select “food” dimension,  AND symbol “&”  and then select  “positive” sentiment.

food & <positive>

Rememeber to “Test Rule” before saving it.

This category will tell you how many said positive about the food in the survey.

Same way you can create the “food & negative” category.


If you wanted to know how many are the negative and positive sentiments related to cities and then use the following:

Bangalore & [<location> +<Negative>]
Bangalore & [<location> +<Positive>]



If you wanted to create location wise airports category, then use the following:

airport & [<location> +<Singapore>]


Please share your text analysis tips and tricks!!!  


Related Post

Related Posts Plugin for WordPress, Blogger...