Have you ever faced difficulty while coming up with an executable idea for a new Data Science project? It surely makes some of us sweat. Many explore datasets to get ideas, which is fun but might take a lot of time and you limit your creativity by referencing!
How can you tackle this problem?
Your goal as a Data Scientist is to strive for creating new and unique sets to generate useful insights and results. So how can you come up with unique ideas? You need to think about only 5 questions in order to figure this out.
These questions will guide your ideation process and enable you to use the full potential of your creativity and in turn, will lead to one of a kind Data Science projects. So let us dive into these questions one by one.
5 Simple steps that will help you ideate better for a new Data Science project
Step-1 Why should you start a project?
Whenever this question is posed, you may answer it by saying that you have a particular goal or intention in mind for the project. But your superior might counter that by asking why you want to do another Data Science project in the first place?
These are queries that help build the process, therefore having a broad categorization of your goal will help you focus on looking for an idea. So, think about whether you want to make:
- A project that allows you to practice a skill, such as natural language processing, data visualization, data wrangling, or a particular machine learning algorithm.
- A portfolio project to show to potential employers.
- A blog post on things such as a concept, model, or an exploratory data analysis.
- Or something completely different.
Step-2 What are your areas of expertise?
This question will help you figure out 3 crucial things. Firstly, as per the Drew Conway’s Venn diagrams of Data Science, being an expert of one of the domains is very important that every Data Scientist should have.
One can only solve data related problems if they understand the underlying problem. But if one is unsure, they end up taking irrelevant steps like creating visualisations, implementing algorithms and making predictions. Why should you do these if they are irrelevant?
Secondly, it is really important to be interested in the idea and dataset you are dealing with. Do not force yourself into doing a project which you are not fully into. If you are interested in an area, you do not need to be an expert in this domain. But you will need to spend time doing extra research and understand the problem further from the data.
Lastly, as per researchers introducing some constraints in the creative process help yielding better results. What this means is that focusing on your areas of interest and expertise will yield better results while making sure if your idea can be turned into an executable idea.
Step-3 How can you find inspiration?
The best option to find inspiration is to read. Many sources can help you identify interesting topics during ideation. You can check out blogs, news, opinion pieces, scientific papers or Data Science posts.
Other than this, you can follow these points:
- Try and combine existing ideas to create a new outcome.
- Explore an existing idea and search for new problems to be solved.
- Take an existing idea and tweak it which completely changes its meaning.
PinStep-4 Where can you find relevant data?
Once you have decided on a general idea, look for datasets to see how you will be able to implement the idea. This will help you determine if your ideas are executable or not. You can check out:
- Existing Dataset sources: AWS, Kaggle, Google Datasets, Buzzfeed, data.world, etc.
- Datasets that other people used: Google your final topic and see if someone has already investigated a similar question. Which dataset have they used? You might get some results on Google scholar as well.
- Collect Data via: Text mining, Web scraping, APIs, log collection etc.
If you are unable to find any data that will help you with your idea, try rephrasing the idea. Rephrasing the idea in lines of the existing dataset is a way to tackle this.
Step-5 Is your idea implementable?
After a lot of hardworking effort, you have a solid idea in hand, but is it ready for execution? Think about all the above steps and recheck that your idea follows them properly. The only thing you will need to determine is if you have the skills to implement and achieve your goal.
Make sure to decide on the amount of time you are willing to spend on the project. Your final project might just be a part of your idea. Therefore:
- Manage all expectations.
- Communicate your idea to a colleague or a senior.
- Be comfortable with starting over.
These steps will surely help you with formulating unique Data Science projects. If you would like to read more blogs like this, check here.
