I did my summer internship at Polaris India Pvt Ltd, an automobile company based in US which manufactures Off-road vehicles. During my internship, i suggested couple of ideas to them and one of them was to analyse the social media presence and come out with some insights. That time I wasnt aware of Gephi and other tools. But now i know how to work on these tools, what can stop me from doing my bit of analysis and reporting them.
There started my Research on Data Analysis of overall Facebook activity on Polaris India Pvt. Ltd. Page and developing business implications based on the analysis.
I made a report and submitted it to my MD, Polaris and also the product manager. They have really appreciated my efforts and are currently discussing the recommendations with the marketing team.
Basic Overview:
Research objective: The main objective of this study is to understand the user’s activity on the FB page, observe patterns if any and finally come out with some business implications based on the analysis. This could be used to segment the Facebook users and understand their preferences, interests etc. so that it:
- Helps us in reaching more people and understanding the customers better
- Also, makes it easy to advertise the region specific events to the target audience which would finally result in getting a better response.
Details of the Dataset:
For analysis purposes I have downloaded the data of the top 50 posts on the page. The total activity associated with these posts which includes likes, comments and shares by users was summing up to 1684. The total number of users who have done such activity is equal to 1025.
The dataset consists of 1075 nodes. These nodes represent the total number of users plus the number of posts (1025 + 50) i.e. whenever there is an activity on the page, Netvizz records it as a node in order to represent the entire activity in the form of a diagram. (For Ex: Consider a case when Polaris Admin posts a picture on FB page and a user likes it. Here, the post becomes one node and the user shows up as another node. The line drawn between both these nodes represent the activity (in this case a “like”).
Initial View of the data:
The bubbles in the below figure represent nodes (Both Admin as well as User’s posts and comments)
The lines are called edges which represent an activity between two nodes. I have used a tool, Gephi, for analysis purposes.
Gephi is an open-source software used for network visualization and analysis. It helps data analysts to intuitively reveal patterns and trends, highlight outliers and tells stories with their data. It uses a 3D render engine to display large graphs in real-time and to speed up the exploration. Gephi combines built-in functionalities and flexible architecture to: explore, analyse, spatialize, filter, cluterize, manipulate, export, all types of networks.
It is designed to follow the chain of a case study, from data file to nice printable maps.
A closer view:
The top three users who have done the maximum activity on the page are as follows:
Some Interesting observations:
- About 85% of the users have activity less than 5.
- 75% of the page activity was generated by photo posts. The response was higher if there was a celebrity mention in the post.
- Status messages by Polaris Admin have generated very less activity and are not posted on a consistent basis.
- Most of the Photo posts didn’t have any messages or titles. Adding text would attract more attention and result in activity.
- Photos which had celebrities had more response (posts like Mr Bhuvaneshwar kumar had 162 likes and 5 shares.)
Filtering the data:
Above figure shows the dataset which is ranked based on activity (i.e. the bigger circles represent nodes with higher activity). Further the data is partitioned based on region and each cluster is coloured differently i.e. the red bubble represents users from UK, green bubbles represents users from US.
The below layout represents the dataset which is segregated based on the activity where the bigger circle represents a post and the smaller circles around it represents the users who are associated with it (either liked or shared it).
A closer view of a cluster is as shown below:
Here:
- Bigger circle represents a wall post done by Polaris page Admin
- The smaller circles represent users:
o Users who are associated with single post
o Users associated with multiple posts
- The screenshot provides information about the total activity and the user location etc.
Summary on Status messages:
The following table shows the details of the status messages along with the total activity associated with them:
Observations:
- Status messages by Polaris Admin have generated very less activity (in this case 9 likes out of total users 106,471 present on the page)
- Status messages by users have very less activity
Recommendations:
- Polaris Admin should post status messages in the prime time (Ex: from 6:00pm to 9:00pm) when more users would be online. This would have more chance of users responding to the posts.
- Polaris Admin must regularly post status messages to engage users.
- Admin must manage to give instantaneous replies to the statuses of users. This would help in creating loyal customer base. (Ex: Status 2 was posted by a user about his experience on 19th July and he got a reply on September 24th due to which there was no more activity.
Picture 1: Number of Likes and Comments for the Status messages on the FB page
Summary on Photo posts:
The following table shows the details of the Photos posted by Polaris Admin and the total activity that was generated:
Observations:
- Photos posted in the prime time had more activity and resulted in more likes and shares.
- Total number of likes generated for the photo posts is 1466 and the average activity = 54.5
- Frequency of photos posts is less i.e. Average number of posts per month is 3.5
- The graph below shows the total activity generated by the photo posts. X axis is the date on which a photo is posted and Y axis is the total activity. We can see a decline in the total activity.
Recommendations:
- Most of the Photo posts didn’t have any messages or titles. Adding text would attract more attention and result in activity.
- Photos which had celebrities had more response ( posts like Mr Bhuvaneshwar kumar had 162 likes and 5 shares.)
Summary on Video posts:
The following table shows the details of the Videos posted by Polaris Admin and the total activity that was generated:
Observations:
- The posts that have celebrity mentions or that are inclined towards some activity have the maximum responses (Likes and shares). Below figure shows the total likes and the number of shares for the video posts.
Few examples are as follows:
- Polaris India - Product Review done by Kranti Sambhav form NDTV India Raftar.Click here for more..... http://www.youtube.com/watch?v=-kfWzZOe3qA
- Polaris RZR 170 in Judwaa Raja .It is the world_s only Side X Side for Youth. http://www.youtube.com/watch?v=3AsXeqAbPDM
- Polaris vehicles are being used around the world for various military operations. http://www.youtube.com/watch?v=wKKeTeRRNBg
Limitations of study:
Since the data extraction was done using my FB account (which does not have Admin access/rights) I wasn’t able to extract the entire information, which includes user’s name, demographics and other interests. This would have helped me done a better analysis.
Conclusion:
Over all, the study has resulted in some interesting insights and recommendations which could surely improve the activity on the page and create more loyal customers. Also, through similar analysis, we could easily segment the users (active and inactive) and target region specific events accordingly so that it would result in a better response and turn-up.
Finally, by creating a mobile app which integrates all these activities, we could easily create, modify these reports on our mobile phone.
There started my Research on Data Analysis of overall Facebook activity on Polaris India Pvt. Ltd. Page and developing business implications based on the analysis.
I made a report and submitted it to my MD, Polaris and also the product manager. They have really appreciated my efforts and are currently discussing the recommendations with the marketing team.
Basic Overview:
Research objective: The main objective of this study is to understand the user’s activity on the FB page, observe patterns if any and finally come out with some business implications based on the analysis. This could be used to segment the Facebook users and understand their preferences, interests etc. so that it:
- Helps us in reaching more people and understanding the customers better
- Also, makes it easy to advertise the region specific events to the target audience which would finally result in getting a better response.
Details of the Dataset:
No of Nodes
|
No of Edges
|
No of Users
|
No of posts extracted
|
Total Activity (Likes and
Comments)
|
1075
|
1671
|
1025
|
50
|
1684
|
For analysis purposes I have downloaded the data of the top 50 posts on the page. The total activity associated with these posts which includes likes, comments and shares by users was summing up to 1684. The total number of users who have done such activity is equal to 1025.
The dataset consists of 1075 nodes. These nodes represent the total number of users plus the number of posts (1025 + 50) i.e. whenever there is an activity on the page, Netvizz records it as a node in order to represent the entire activity in the form of a diagram. (For Ex: Consider a case when Polaris Admin posts a picture on FB page and a user likes it. Here, the post becomes one node and the user shows up as another node. The line drawn between both these nodes represent the activity (in this case a “like”).
Initial View of the data:
The bubbles in the below figure represent nodes (Both Admin as well as User’s posts and comments)
The lines are called edges which represent an activity between two nodes. I have used a tool, Gephi, for analysis purposes.
Gephi is an open-source software used for network visualization and analysis. It helps data analysts to intuitively reveal patterns and trends, highlight outliers and tells stories with their data. It uses a 3D render engine to display large graphs in real-time and to speed up the exploration. Gephi combines built-in functionalities and flexible architecture to: explore, analyse, spatialize, filter, cluterize, manipulate, export, all types of networks.
It is designed to follow the chain of a case study, from data file to nice printable maps.
A closer view:
The top three users who have done the maximum activity on the page are as follows:
Name
|
Place
|
Gender
|
Total Activity
|
George Enotiades
|
el_GR (Cyprus)
|
male
|
20
|
Pamesh rawat
|
en_US (Kulu, HP)
|
male
|
18
|
Om Patel
|
en_US (Nagpur, MP)
|
male
|
18
|
Some Interesting observations:
- About 85% of the users have activity less than 5.
- 75% of the page activity was generated by photo posts. The response was higher if there was a celebrity mention in the post.
- Status messages by Polaris Admin have generated very less activity and are not posted on a consistent basis.
- Most of the Photo posts didn’t have any messages or titles. Adding text would attract more attention and result in activity.
- Photos which had celebrities had more response (posts like Mr Bhuvaneshwar kumar had 162 likes and 5 shares.)
Filtering the data:
Above figure shows the dataset which is ranked based on activity (i.e. the bigger circles represent nodes with higher activity). Further the data is partitioned based on region and each cluster is coloured differently i.e. the red bubble represents users from UK, green bubbles represents users from US.
The below layout represents the dataset which is segregated based on the activity where the bigger circle represents a post and the smaller circles around it represents the users who are associated with it (either liked or shared it).
A closer view of a cluster is as shown below:
Here:
- Bigger circle represents a wall post done by Polaris page Admin
- The smaller circles represent users:
o Users who are associated with single post
o Users associated with multiple posts
- The screenshot provides information about the total activity and the user location etc.
Summary on Status messages:
The following table shows the details of the status messages along with the total activity associated with them:
Status
|
Title
|
Likes
|
Comments
|
Total Activity
|
Status 1
|
I like polaris 4×4
|
0
|
0
|
0
|
Status 2
|
I just had ride it yesterday
|
0
|
1
|
1
|
Status 3
|
Come and Enjoy the free Test Ride at our Lucknow Dealership Launch on
20 July 2013 and 21 July 2013.For details please go through our event
section.
|
9
|
9
|
9
|
Observations:
- Status messages by Polaris Admin have generated very less activity (in this case 9 likes out of total users 106,471 present on the page)
- Status messages by users have very less activity
Recommendations:
- Polaris Admin should post status messages in the prime time (Ex: from 6:00pm to 9:00pm) when more users would be online. This would have more chance of users responding to the posts.
- Polaris Admin must regularly post status messages to engage users.
- Admin must manage to give instantaneous replies to the statuses of users. This would help in creating loyal customer base. (Ex: Status 2 was posted by a user about his experience on 19th July and he got a reply on September 24th due to which there was no more activity.
Picture 1: Number of Likes and Comments for the Status messages on the FB page
Summary on Photo posts:
The following table shows the details of the Photos posted by Polaris Admin and the total activity that was generated:
- Photos posted in the prime time had more activity and resulted in more likes and shares.
Timestamp
|
Likes
|
Shares
|
2013-07-19T18:05:39+0000
|
139
|
4
|
2013-07-19T18:00:39+0000
|
162
|
6
|
2013-07-19T17:54:20+0000
|
151
|
5
|
- Total number of likes generated for the photo posts is 1466 and the average activity = 54.5
- Frequency of photos posts is less i.e. Average number of posts per month is 3.5
- The graph below shows the total activity generated by the photo posts. X axis is the date on which a photo is posted and Y axis is the total activity. We can see a decline in the total activity.
Recommendations:
- Most of the Photo posts didn’t have any messages or titles. Adding text would attract more attention and result in activity.
- Photos which had celebrities had more response ( posts like Mr Bhuvaneshwar kumar had 162 likes and 5 shares.)
Summary on Video posts:
The following table shows the details of the Videos posted by Polaris Admin and the total activity that was generated:
Title
|
Video
|
Likes
|
Shares
|
Polaris India - Product Review done by Kranti
Sambhav form NDTV India Raftar.Click here for more.....
http://www.youtube.com/watch?v=-kfWzZOe3qA
|
Video 1
|
6
|
3
|
Polaris India Pvt. Ltd shared a link.
|
Video 2
|
3
|
|
Polaris India Pvt. Ltd shared a link.
|
Video 3
|
3
|
1
|
Polaris India Pvt. Ltd shared Polaris Goa - Kunde
Cars & Bikes Pvt. Ltd._s video: Polaris ATVs on track.
|
Video 4
|
6
|
|
Polaris India Pvt. Ltd shared a link.
|
Video 5
|
15
|
|
Polaris India Pvt. Ltd shared a link.
|
Video 6
|
7
|
1
|
Polaris India Pvt. Ltd shared a link.
|
Video 7
|
18
|
|
watch RZR S 800 and Sportsman 500 HO Touring in
motion https://www.youtube.com/watch?v=6WrdpSN-Kdc
|
Video 8
|
16
|
|
Polaris India has a wide Range of ATV and Rangers
starts from 50cc to 900cc.Watch this for more.
https://www.youtube.com/watch?v=ZIsOCufVgF8
|
Video 9
|
8
|
|
http://www.youtube.com/watch?v=WK4x6pP0rs8
|
Video 10
|
12
|
|
Polaris RZR 170 in Judwaa Raja .It is the world_s
only Side X Side for Youth. http://www.youtube.com/watch?v=3AsXeqAbPDM
|
Video 11
|
21
|
4
|
Polaris vehicles are being used around the world
for various military operations. http://www.youtube.com/watch?v=wKKeTeRRNBg
|
Video 12
|
8
|
2
|
Observations:
- The posts that have celebrity mentions or that are inclined towards some activity have the maximum responses (Likes and shares). Below figure shows the total likes and the number of shares for the video posts.
Few examples are as follows:
- Polaris India - Product Review done by Kranti Sambhav form NDTV India Raftar.Click here for more..... http://www.youtube.com/watch?v=-kfWzZOe3qA
- Polaris RZR 170 in Judwaa Raja .It is the world_s only Side X Side for Youth. http://www.youtube.com/watch?v=3AsXeqAbPDM
- Polaris vehicles are being used around the world for various military operations. http://www.youtube.com/watch?v=wKKeTeRRNBg
Limitations of study:
Since the data extraction was done using my FB account (which does not have Admin access/rights) I wasn’t able to extract the entire information, which includes user’s name, demographics and other interests. This would have helped me done a better analysis.
Conclusion:
Over all, the study has resulted in some interesting insights and recommendations which could surely improve the activity on the page and create more loyal customers. Also, through similar analysis, we could easily segment the users (active and inactive) and target region specific events accordingly so that it would result in a better response and turn-up.
Finally, by creating a mobile app which integrates all these activities, we could easily create, modify these reports on our mobile phone.
No comments:
Post a Comment