[Music] Titanic data set!! Number one reason that you're not getting an interview call for a data science position, is poor project portfolio in your resume. In this video i'm going to talk about five essential projects that you need to have on your resume to attract interviewers attention. The first two projects are going to be BI projects which is more suitable for data analyst, although it could be helpful for data scientists as well and the next three projects are going to be machine learning projects, which are for data scientist position. Now in the end of the video, I will discuss some general tips that you can use to generate project ideas. The projects which are very unique and which looks cool on your resume. Let's get started! SQL and Dashboarding in BI tools such as power BI and tableau are two most important skills for a data science career. On my youtube channel there is this project series on in power bi and tableau both for generating sales insights for a business. Now I did this project with the help of my friend Hemanan, who is a data analytics manager. So this project was done in a corporate style, where you start with a problem statement, then you do project planning then data discovery, then you do simple analysis using sql and then you do the same analysis by pulling that data into power bi doing some data cleaning, etl building dashboards, then you take feedback from customers and then you build a version 2. So this was a very practical corporate style project series that I have done and in fact, I have received so many testimonials where people got a job or internship just based on this project. So you can use this project almost like a seed idea and you can build maybe your own unique project and do dashboarding in tableau or power bi. To customize this project you can use the data sets from Kaggle or Nasdaq Data Link. There are so many data set websites out there, so just get data from there and do the similar kind of dashboarding. Today's video is sponsored by Udemy, which is one of the top e-learning platform. Udemy offers courses on diverse set of topics including machine learning data, science python, and so on. Let's say I'm interested in learning power bi. You can search for power bi and get so many courses. Now the best thing here is the reviews. So if you go to this particular course, for example it has been taken by 166000 students, 4.6 star rating. You can read the ratings and you can't go wrong. You know these courses offer so much valuable content at a very reasonable price. Try the link in the video description below to get up to 80% off on your udemy course purchase. Back to our video now. Baby yoda is asking me about my home expenses, my rent, my groceries, and so on. And to answer his questions I built a personal finance dashboard in power bi, again. Yes baby, I have answers buddy for your questions and using this personal finance dashboard you can get answers on your personal finances. So in a way you are kind of your own client and I did this project with the help of my data analyst friend Hemanan uh and this was again uh like a real life project where you are getting answers of your own expenses and building various dashboards, which can be pretty useful. So this project series is also available on my youtube channel, so you can check it out. You can build maybe an excel file for your or your own expenses and you can draw various kind of charts. Next three projects are going to be machine learning projects. When you talk about machine learning you want to have projects, which covers diverse set of topics such as classification, regression, deep learning. In deep learning again nlp and computer vision and then you can have a project on clustering as well. For classification I have this project series on spot celebrity image classification where we build an end-to-end website that can classify the spot celebrity. Now you can use this project as a seed idea and maybe you can build a project for movie star classification or spiritual guru or you know political leaders or entrepreneurs who are your most favorite, you can classify those personality or even better you can classify your own family member pictures as well. So the idea is you look at this project on my youtube channel. It's like few video series where i used fat [ __ ] chrome extension for image data collection. Then I used some of the data cleaning and pre-processing techniques then build a model using support vector machine. Usually for image classification I know, you use cnn but in this case I used support vector machine and then we did some optimizations grid search cv for uh hyper parameter tuning and then the model deployment. And then the model was being run inside a flask server, which is a python web server and then the ui was simple html html css javascript. So end to end project you build the entire application, you learn a lot of useful skills, and you use this project as a reference to learn the project architecture technical skills uh but then you customize the project either for political leader classification, or your family member classification and when you put that project in your resume it's gonna look very unique. The fourth project I suggest you add to your portfolio is going to be a regression project. On my youtube channel I have this bangalore property price prediction project where I build an end-to-end website, where you can enter your parameters you know your bedrooms, your square foot area, and so on and the website will tell you the predicted price for that home. Now I know you you might be thinking home price prediction is kind of old boring problem. There is nothing unique about it, but you can make it unique by maybe doing a property price prediction in New York or Bangkok different city or maybe take a different uh problem altogether, related to regression. Let's say you love cricket and the IPL- Indian Premier League is playing, so you want to do a prediction on who is going to win the next IPL match or how much Virat Kohli is going to score in the next match. So these are regression problems and you can get all the required data from internet maybe explore Kaggle or IPL websites and get that data. Now when you do something that you love that you are passionate about, let's say you're passionate about cricket then do IPL sports analytics. You're passionate about gardening or let's say music or even environment then do some project on you know how the pollution is changing the environment and things like that. So you are doing something that you are passionate about it keeps you motivated and then that project looks pretty unique on your resume. Going back to your my bangalore property price prediction project um I built this website where uh the model was trained uh using simple linear regulation and before training the model we did outlier removal, we did dimensionality reduction, a lot of feature engineering techniques then we built the model optimized it hyper parameter tuning using grid search cv. Then we deployed the model to amazon aws cloud and there was a flash server running on top of it, which was serving the request that were coming from the website. So this way you build the entire solution you learn a lot of technical skills by going through this complete video series and by the way this is all free. All the code is available on github. So if you have a learning spirit you can learn things pretty easily uh using these videos and then you're customizing it, you're adding your own flavor and when you put that project on your resume, you will most likely get an interview call. I grow nice tomatoes in my backyard and let me show you how they look and sometimes the problem that I face is they get some disease and to detect the disease in the plant I built an application in react native and used deep learning behind the scenes. Y+This mobile app that was built can take a picture of a plant and it can immediately tell you what kind of disease this plant has. Now i used potato disease data, so it was the if the application was built for potato disease classification but you can use it to build a classification for decisions in tomato plant or maybe take bunch of MRI scans of brain to detect if the brain has a tumor or not. So again the idea is to give you the whole project structure, but then you can customize it and give it your own flavor. If you go to youtube and search for code basics potato disease classification project, you will find my playlist and in that playlist you will see videos starting from problem statement to data collection model building fast api server and so on and just to go over a technical architecture. We first perform data cleaning and pre-processing that where use tf data set, data augmentation and then we do model building using convolutional neural network and then we export the model into tf lite model and tf serving. We first build a website and then we you know build a mobile application as well, where the tf lite model is exported as a Google cloud function. so you're learning a lot of useful cloud tips as well and then you build this react native app which makes calls to this Google cloud functions. Again this is an end-to-end project where you can build the mobile application for brain tumor detection for any kind of other disease detection or in general any kind of image classification. So those were five projects that you need to have on your resume for data science career. Now let's look at three tips that anyone can use to generate unique project ideas. The first tip is something we already discussed, which is look at the internet, look at my youtube channel or other youtube channels to get the seed ideas and then customize them and come up with your own unique flavor. So that's one. Then you can also use a famous freelancing website Upwork to get the machine learning or data science project ideas. In Google search upwork machine learning projects and you can go to upwork.com, which is a freelancing website and find all the machine learning projects here. For example road sign recognization system. Now these are the real problems posted by real clients or businesses and these problems reading through these problems will give you a lot of seed ideas. You will not find any data set here unless your contract is signed but the idea is you at least get the ideas for your projects. Number two is helping someone for free. There are so many open source projects available on internet where you can collaborate with smartest people in the world and you can build a solid project portfolio. Open course contributions have a huge value on your resume. So they not only make your project portfolio stronger but you build a lot of connections and it can help you find job pretty easily. You can also help an NGO or if you have any relative who have a business, you can offer your help to them for free and in that pursuit you will learn some important skills and you'll be working on real life projects you know, which looks pretty good on your resume. Tip number three is look at the websites that that makes different data sets available and by looking at the data sets sometimes you can get project ideas. The popular websites for data data sets are Nasdaq Data Link, Kaggle, Google Data Set search, etc and when you go through various data sets on these websites, just by looking at the data sets you will get a project idea and if you don't uh join my discord server. I have a project channel on my discord server, where you can brainstorm project ideas with other folks. So sometimes you know doing project alone might not be that effective but if you do it in a group you get some novel ideas and it will be very very helpful overall. That's all I had. If you like this video please give it a thumbs up and if you have a unique data science project idea please post in a video comment below. Thank You! you
