Malta Elections 2017
The Social Activity Scoreboard
Malta Elections 2017
Welcome to ICON and Minely’s Social Activity Scoreboard. During the election period, we carried out a detailed analysis on social media campaigns carried out by the major parties in the run-up to the June 3rd elections. Combining Minely’s big data and visualisation skills with ICON’s deep understanding of the social media space, we’ll be providing you with a unique perspective on the elections.
In the ‘Dashboard’ section, you’ll find a number of charts outlining sentiment on the different parties based on the public’s liking of their Facebook posts. We’ve also analysed the distribution of PN vs PL sympathies in user activity across the major local portals. Finally, there’s a list of posts issued by different media and the relative success of the post based on user engagement.
Raw data’s well and good, but we’ve taken it a step further by creating a detailed analysis with our own insights in the form of a report. If you’d like to find out more about big data or social media analysis and how these can help accelerate your business, click here.
Social Media Analysis
This report contains insights of the 2017 General Election Political Campaigns communicated on Dr Joseph Muscat’s and Dr Simon Busuttil’s Facebook pages. Both political parties have their own pages; Partit Laburista (PL) and Partit Nazzjonalista (PN) respectively. However, since the leaders of both parties have significantly larger numbers attributed to their personal pages, the analysis focuses solely on these.
This section contains an overview of the methodology employed to collect, analyse and process the data for the purpose of generating this graph. The activity is done in steps as follows:
We created a Facebook App to access its Audience Network and downloaded public posts and reactions for the last 90 days. A Facebook App is an interactive software application developed to utilise the core technologies of the Facebook platform.
We then analysed all the social reactions from Simon Busuttil and Joseph Muscat’s Facebook pages. This allowed us to identify all the user-IDs who liked a post (but not the actual followers). We were then able to create three research personas:
- PL Absolutes: These are users who only ‘like’ PL posts in exclusion of PN posts.
- PN Absolutes: These are users who only ‘like’ PN posts in exclusion of PL posts.
- Sympathisers of both parties: These are users who occasionally ‘like’ posts from both parties. This last category is critical as it defines what is traditionally known as the ‘switch’ vote.
For each post published on social channels, we created the following datasets using social-reactions available:
- The number of users who liked a post and are also ‘PL Absolute’ users
- The number of users who liked a post and are also ‘PN Absolute’ users
- The number of users who liked the post and are ‘Sympathisers of both parties’
- The share count
- The like count
- The emotion status based on ‘love’ count, ‘sad’ count, ‘angry’ count, ‘wow’ count and ‘haha’ count
Features 1, 2, and 3 help us determine whether the post is pro-PN, pro-PL or neutral and we apply k-means machine learning algorithms to cluster the posts in these 3 segments . Features 4 to 6 help us evaluate non-political posts. We identified a pattern that political posts tend to have positive and negative reactions (for example Like and Angry). Posts of non-political subjects tend to be either mostly positive or mostly negative.
We took the posts and the features and we trained a machine learning model to cluster the posts in 3 groups: PN/PL/other. Machine learning is a type of Artificial Intelligence (AI) that provides our software with the ability to learn without being explicitly programmed. Machine learning focuses on the development of applications that can react correctly when exposed to new data (such as new posts).
The resulting clusters were then correlated to analyse the sentiment. Sentiment analysis – otherwise known as opinion mining – is the process of determining the emotional tone and user-interest behind post-content used to gain an understanding of the attitudes, opinions and emotions expressed within such online mention.
In terms of accuracy, we have manually analyzed a sample dataset of 500 posts (from the 30,000+ used so far). The method for filtering the ‘noise’ ( i.e. filtering political posts from non-political posts using clustering) is 86.6% accurate. Once a post is classified as political, the classification of PN/PL is 99.9% accurate allowing us to express a high degree of confidence in our findings.
Find Out More
Interested in harnessing the power of big data and social media analytics for your business? Fill in the form below and get in touch with us today.
How do you get this data?
Minely offers a full range of social connectors including Facebook, Twitter, Instagram etc. Using a Facebook Application Id, a connection was made to the social network. Minely retrieves public data periodically from the social pages of each news portal. Using a combination of natural language processing and machine learning data is extracted and analysed into Minely.
Will this help me know who will win the election?
No, this tool cannot predict who will win the general election in Malta. It merely provides an insight into activity and engagement across social channels.
Some of the posts above are not political, why are they here?
We have built a Machine Learning Model which clusters over 25,000 articles and over 2 million reactions automatically. The clustering is used to determine whether a post is non-political or to classify as PL or PN. This model utilises the political parties’ social pages as initial sentiment signal for building the predictive model. The resulting model is used to automatically classify articles from all portals’ including independent media. A machine learning model represents a real-world scenario. It abstracts from the real relationships by simplifying variables which are not too important. This simplification process causes a margin of error which is considered as noise.
Are you accessing my social profile without consent?
No, we do not have access to any personal profiles. Minely can only retrieve public data.
Terms & Conditions
Your use of the data from the ‘Malta Elections 2017 Social Activity Scoreboard’ (referred to as the “Service(s)”) is subject to the terms contained in this document.
The Service is provided by ICON Studios Ltd and Minely Ltd (referred to jointly as the “Service Provider”).
The Service is provided “as is” and Service Provider gives you no warranty with respect to it.
Service Provider has made every attempt to ensure the accuracy and reliability of the information provided in the Service. However, the information is provided “as is” without warranty of any kind. Service Provider does not accept any responsibility or liability for the accuracy, content, completeness, timeliness, legality, or reliability of the information contained in the Service or that such information and data will be error-free.
All warranties of any kind, express or implied, including but not limited to the implied warranties of merchantability, fitness for a particular purpose, freedom from contamination by computer viruses and non-infringement of proprietary rights are disclaimed.
In particular, Service Provider, does not represent or warrant to you that (A) your use of the Service will meet your requirements; (B) your use of the Service will be uninterrupted, timely, secure or free from error; (C) any information obtained by you as a result of your use of the Services will be accurate or reliable; and (D) that defects in the operation or functionality of any Software used to provide the Services will be corrected.
No conditions, warranties or other terms (including any implied terms as to satisfactory quality, fitness for purpose or conformance with description) apply to the Service except to the extent that they are expressly set out in these terms.
You agree that you are solely responsible for any content that you create, transmit or display derived from the Service and for the consequences of your actions by doing so. You shall indemnify and hold Service Provider harmless from all claims and all liabilities, costs, proceedings, damages and expenses awarded against, or incurred or paid by Service Provider as a result of the use of such Service.
This Service has been prepared for general guidance on matters of interest only and does not constitute professional advice. You should not act upon the information contained in this Service without obtaining specific professional advice.
Service Provider, its members, employees and agents do not accept or assume any liability, responsibility or duty of care for any consequences of you or anyone else acting, or refraining from acting, in reliance on the information contained in this Service or for any decision based on it.
You are not permitted to use or distribute the contents of this presentation without the written consent of Service Provider.
© ICON, 2017 © Minely, 2017