Machine Learning is getting a hotter topic by the day. Many talk about it, but few know really what they are talking about. With so many resources and tutorials, it might get pretty overwhelming where to start from if you have no experience with the matter...

In case you have played with the idea of making a Natural Language Processing project of your own, we have just the right tips to share with you so you can kick it off on the right foot. Read on!

The Experiment

One of our junior engineers Dave recently became more and more fascinated by Machine Learning and really pondered on exploring this new and exciting field. Last week he decided to start a small project involving text by using the most common natural language processing and machine learning techniques to see where this would take him.

Like you can imagine, as a total newbie in this field - he had to do a lot of reading and watching tutorials. :)

The Resources

For a couple of days, Dave had read tons of tutorials and watched numerous videos on the topic until he stumbled on a YouTube lesson from Alice Zhao on Natural Language Processing in Python.

In contrast with all he had gone through, this recording was more focused on Data science workflow rather than on the coding part, libraries, etc. And suddenly it all made sense! Dave’s feedback after watching summarizes why you need to watch this video first: “It would have saved me hours reading and banging my head on the wall if I had found this on day one!”

What is Data Science Workflow?

Data science workflow refers to a process or iterative set of steps for solving cases and addressing complex data problems like:

• Text pre-processing (Document-Term Matrix)
• Exploratory Data Analysis (visualizations and stuff)
• Sentiment Analysis (TextBlob)
• Topic Modeling (Latent Dirichlet Allocation / nltk, gensim)
• Text Generation (Markov Chains)

In short, Alice’s lecture shows a very clear, practical, and funny way to begin and complete an NLP project from scratch - just watch it and don’t be afraid to try new things.

Conclusion

Dave began his journey without having a clue about where to start. Thanks to him sharing his story, you don’t have to!

I would also like to share that all popular NodeJS Machine Learning technologies on NPM are supported on www.SashiDo.io. On top of that, only until the 31st of May, you can take advantage of our extended 45-day free trial to begin your Machine Learning journey unrestricted and start building Smart APIs.

Dave and the team will be happy to give you some hints on how to tackle any challenges you might face. ;)