UI/UX Study Case: Sentiment Analysis Dashboard
In this case study, I explain my process in creating a Sentiment Analysis Dashboard Design for one of Indonesia Health Organisations from the first step (brainstorming process) to the final design.
⚠️ Disclaimer : The project is handled by Venation and has obtained publication permission. I am part of Venation which is working on this project directly.
📝 OVERVIEW
Background
Our client is one of Health Organisation in Indonesia, the company is responsible for taking care of domestic health affairs. They also commit for formulating, establishing, implementing, managing and supervising all health activities in Indonesia.
👨🔬 RESEARCH
Background Research
The client‘s team wanted to create a live dashboard regarding the sentiments which are being discussed or popular in the public society.
Brainstorming
In this project, I’m as an UI/UX Designer and the Developer team do the brainstorming so that the dashboard could run the client’s expectations. The results of the brainstorming are as follows:
- What tools will be used in working on this dashboard?
- Tableau
- Visual Studio Code
- Figma - What features/cards will be displayed in the dashboard?
Twitter and News Portal as an analysed objects because Twitter is one of the text platforms that has the most users in the world especially in Indonesia, and News Portal is a source of reports on facts and showing what is trending.
— Twitter Dashboard
- Topics, client provides several topics that wanted to be analysed so that the appears of data could be more specific and directed.
- Location, aims to find out the spread of sentiment in every province in Indonesia.
- Sentiment, here we use Pie Chart to visualise the total number of positive, neutral, and negative sentiments of each topics.
- Gender, here we use Pie Chart to visualise the total number of Male and Female.
- Tweet Overtime, here we use a Line Chart to visualise tweet sentiment with the latest time range.
- Word Cloud, here we use Word Cloud Chart to determine which is the most discussed words.
- Summary, here we use a Bar Chart to visualise the total numbers of sentiments based on the most frequently occurring words.
- Positive, Neutral, and Negative Sentiments, here we use a Chart Table to visualise the sentiment results containing Username, Tweet, Location and Retweet Count.
— User Page Dashboard
Contains the profile of the account that made the tweet in the form of Name, Username, Location, Date of Joining Twitter, Number of Following and Followers, Twitter Account Link, Tags, Tweets Detected, and Issue Participated (Sentiment).
— News Dashboard
- News, here we use Chart Tables to visualise sentiments containing News Publishing Dates, News Portals, News Headlines, and Sentiments.
- Sentiment, here we use a Text Chart to visualise the total number of Positive, Neutral and Negative sentiments.
- Tweet Overtime, here we use a Line Chart to visualise news sentiment in the latest timeframe.
- Word Cloud, here we use Word Cloud to determine the most words which used as a news headlines.
My Roles
User Persona
Users who will use this dashboard are internal employees of the client’s company.
Goal
Create a Dashboard which able to visualise data live and easy to use.
🛠 WIREFRAMING
User Flow
I used the User Flow method to create an overview of the stages that describe user interactions with the dashboard they will use.
Sketch
I made a simple and quick sketch to visualise the Dashboard appearance as discussed with the Developer team. This stage helps me to explore design ideas more interesting and insightful.
Low Fidelity (Lo-fi Design)
I applied images from Sketch into a mockup design that focused on Dashboard features, flows, and layouts.
At this step I make sure the features and dashboard layout have been approved by the client before moving on to the next step
🎨 DESIGN
Design System
After the features, flows, and layouts of the Dashboard were approved, I created a Design System to categorise the design according to its function into several component forms. The Design System makes the design consistent and flexible.
High Fidelity (Hi-fi Design)
After the Design System is complete, I put and integrate these components into the dashboard according to the layout that I made in the Lo-fi Design step. This step is the last part in the design process.
🖼 PROTOTYPING
At this stage, I created a clickable prototype using Figma. This prototype contains 3 dashboard pages. I use this prototype to be presented to the client, after the client is approved, the Developer team will do their job to make this dashboard design live.
🔬 TESTING
Usability Testing
Usability Testing is the dashboard testing phase by the client which aims to ensure that the dashboard is in accordance with the client’s requirements in terms of functionality and user interface. So this dashboard can be used smoothly. This testing is done to get feedback from clients.
Feedback
- The number of words in the Word Cloud feature needs to be adjustable.
- Need a range of data retrieval can be customised.
💡 Conclusion
Some of the things that I learned in working on this project, such as communicating well with clients and developers, prioritising urgent work, and good time management.
“The best design is a design that is approved by the client and can be done by the developer” — Andi Muh Shabrani
“Sebagus-bagusnya desain adalah desain yang diterima oleh client dan bisa dikerjakan Developer” — Andi Muh Shabrani