Freelance Data Scientist & Statistical Consultant: Interview with Wesley Engers
Wesley Engers is a highly successful and sought after Freelance Data Scientist & Statistical Consultant. He expertise in providing data insights to companies and consultants using applied mathematics, statistics, and human ingenuity and has good exposure using SPC Excel Add-in, SPSS, SQL, R, Rshiny, SAS JMP etc. In a chat with Kalyan Banga, Wesley gives a great insight into the world of freelance data scientist.
- Hi, Wesley! Can you tell us a little bit more about yourself?
Sure, I am a freelance data scientist and math tutor. I have always enjoyed working with numbers and seeing what they can say about the costs and benefits of a given situation. Currently I use freelance websites (such as Upwork and Experfy) and referrals to find clients who have need of my data analytic services. I use my extensive math and statistical knowledge as well as software (generally R) to advise clients on their questions. In the past, I have worked for Symantec (Fortune 500) as an Associate Statistician doing data analytics in both statistical process control as well as marketing, pricing, and discounting analytics.
- What is your academic background? Which is the most important aspect of your expertise?
I have a Bachelors degree in Math and Finance from Santa Clara University and Masters degree in Mathematics from Claremont Graduate School. Having a deep knowledge of mathematics and statistics is at the core of my business and work. I use it extensively to figure out the most appropriate models to use in any given situation as well as how to apply various mathematical concepts to new situations (creativity is very important to finding informative solutions). Additionally, I try to expand my knowledge by taking online classes through Coursera.
- What prompted you to getting into freelancing? How did you get your first freelancing work in analytics? You have team or you work alone?
I’m a big believe in having multiple income streams being a safer bet that one large income source. So seeking out these income streams led me to freelance in addition to my love of problem solving.
I got my first freelancing work from Elance which later merged with ODesk to become Upwork. At that point, I was really trying to establish myself and tended to bid fairly low to help build up my reputation.
I currently work by myself on the work I do. However, I do have other consultancies (such as EdgePeak, LLC.) that I work with/through. They generally have other clients that need data analytics work completed so I’ll be the go to person for that aspect of a larger project.
- What are some of the advantages of freelancing vs. a full-time analytics position?
Freelancing provides a lot more flexibility than full time work but that also comes with a lot more responsibility. A freelancer has the flexibility to set their own schedule and work however many hours they want. I could work say only 20 hours a week or less and work at very odd hours such as 10pm to 2am if I wanted. However, it is also the responsibility of the freelancer to find that work, market themselves, and then have the discipline to do a high quality job on the work they committed to. To meet those responsibilities sometime you may have to work a bit when you originally planned to be on vacation or take some phone calls at early hours because your client is half-way around the world. It’s definitely not for everyone, but it will be fantastic for those with the right mindset.
- How your typical day does looks like? How do you juggle your freelancing work along with consulting or other engagements? If you can share details of few freelance projects.
I don’t think I have a “typical day” because I’ve tried to build my schedule around maximal flexibility. However, I prefer to sleep in till around 9 or 10am then read news (keep up on current happens in both the real world and data science content) and emails. I’ll probably shower and might get some late morning/early afternoon work done then grab a late lunch. During the school year I am often doing math tutoring during the afternoons and evenings or perhaps more freelancing work during the summers. Often an evening workout will follow every other day and grab some dinner. Finally, get some more work done late night into the early morning and go to sleep. Work in this case will include both actual projects I’m working on and also work on getting additional projects and work for the future.
Here’s a link to my Upwork profile, where I have recorded much of my online work. I personally tend to focus on the “decision science” side of data science and much less on the “data product” side of things. Hence a typical project for me would involve helping to make business decisions such as “which customers are likely to churn?”, “what are our typical customer profiles?”, “what do our best customers look like?”, or “who should we target to grow our userbase?”. Therefore, not only do I need to do data analysis and the math side of things I also need to communicate that to stakeholders in a manner and language that they understand and can take action on.
One recent project involved predicting auction prices for particular collectables. Very tricky given the wide variety of products and the uncertainty of pricing in an auction setting. Basis of the analysis was regression but some interesting transformations were needed to get the data into a good format so that it was usable. Lots of a data scientist time is spent cleaning and getting data into a usable format so this is very typical. Once the model was built then refinements were done to fine tune the model to make the best predictions.
You can also view some of the work I’ve done for Coursera classes on my Rpubs page. This is fairly typical of the work I do but haven’t really added in the refinements on the presentation side or the fine tuning of models as that isn’t essential for passing courses. New content is great for thinking of ideas that you can combine to create innovative solutions to problems.
- What are the steps one needs to take to find freelance data scientist work for newbies?
First, you need to have the technical skills (math, statistics, and programming) and understand so that you are actually providing value to the organization and people you are serving. This can be gained from an academic background, some online courses, and of course experience is always a great teacher. I’m not sure there is any set path to data science or a career but I’ve always found that doing things I find interesting to be a good guide. If I’m interested and passionate about a topic then I’ll be willing to spend the time to do it right even if it doesn’t make the most economical sense at first. And if it doesn’t work out then at least you learned something you valued.
- Any tips for someone who wants to get their first contract? How much one can expect to earn? Compare a fresher to 5 and 10 yr experienced guy.
Newbies should definitely be go-getters. It’s ok to push yourself a little out of your comfort zone as that is how you learn (not too far or you won’t be able to turn in good work). Also, never underestimate the power of following up. Respond in a timely manner to all requests and if you don’t hear back from them follow-up and see where things are at (this goes for both clients that you have as well as prospective clients you’re trying to engage).
Earnings really depends on your skill set and the client’s needs. Could be anywhere from $20 to $200+ per hour minus the standard industry cut of 20%. Maybe, your first project is only $100 or $200 something small to get your feet wet. My strategy was to start small and build up but depending on your skills there isn’t much stopping you from trying to land a bigger project.
- What are the challenges for an aspiring freelance data scientist? How can s/he overcome those? Initially after building a profile, freelancers face difficulty in getting new assignments, thus building credibility to bag further projects are a struggle? Any advice from Upwork experience.
Finding the right client’s is always a challenge because there is so much that you can’t control about it. You can always work on learning new skills and find the right places and people to build those skills but clients are more difficult.
- As mentioned before, always follow-up and be persistent (not annoying) in pursuing potential clients.
- You can also start off by trying to low-ball a project a little bit so that someone will choose you for having a good buyer price but be careful with this strategy as it can backfire if you get caught in the price wars.
- Most important will be to do an incredible job once you get the job. Try to blow your client away every time and anticipate what their next needs are. If you can anticipate their next ask and already have an answer for them, they’ll be very impressed.
- Think outside the box and try to come up with less common solutions to their problem. Most people use common machine learning algorithms (random forest with gradient boosting for example) or regression. Certainly, use these common tools as a building block but usually you need to add a “twist” or innovation of some sort to make your proposal stand out. Also, try to give them a variety of potential solutions with pros and cons of each and let them pick the direction they want to take the project.
- Which websites can aspiring freelancers enrol themselves and how much they should charge typically for per hour service? Can you give a rough estimate of how charges varies based on experience, geography (India vs. say US), etc on freelance websites?
I use Upwork extensively. Experfy is also good as it specializes in Data Science projects. I’ve heard of freelancer.com and toptal.com as well but haven’t tried them out.
Charging varies greatly and could be anywhere from say $20 per hour to $200+ per hour. But I’m definitely not in the business of telling other people how much they should or shouldn’t charge. They should work on figuring out the value of their services and their time and charge accordingly. I started on the lower end to help build up my reputation but you definitely don’t want to get caught in a cycle of charging too little. I think that one of the biggest mistakes made is not charging enough for your services. Also, once you have a better idea of how long a typical project type takes and how valuable it is a fixed price project rather than hourly rate could be beneficial.
- What are the kinds of projects that are offered through freelance mode? Is it more of dashboard, visualization work or modelling, prediction kind of work? What is the typical duration of each such projects?
I’m sure all kinds of projects are out there depending on your interests and skill set. I tend to favour the modelling and prediction type projects that inform decision making.
Typical duration could be anything from 1 to 2 weeks up to 6 months or a 1 year+. It really depends on the scope of the project.
- Which programming languages do you know? Which one is your favourite?
I know R, SQL, some Python, and have academic experience with Matlab and C++. R is definitely my go to programming language and I do most of my work in there.
- Which analytical tools and software’s you use for your freelance and consulting work?
I mostly use R, some SQL and some Excel. Excel is very good because almost all business type people are familiar with it but is certainly limiting in its advanced capabilities (both data handling and advanced analytics). SQL is great for knowing how to pull data and give a good basis for structuring data as well. I use R for the heavy hitting stuff. It can handle large amounts of data (I understand that some CS and data warehouse types might disagree here but for my clientele it’s sufficient) as it is limited by your laptop capabilities and it’s open source so I almost always have access to the latest algorithms and advancements. Additionally, provides good visualization and plotting so I can easily present the results (usually via Power Point but have heard Tableau can be great for presentation) to non-mathy people.
- If you were given a chance to rebuild your analytics career, what changes you would have made to have even better growth and fulfilment?
Very happy where things are and how things have turned out. I don’t think I’d change anything. I think one key is to finding what you want to do and are happy doing. A lot of times people stick around too long in a job they hate just for the money. Always think about the alternatives. It doesn’t have to be as dramatic as quitting your full time job and having no income but just spending a few hours per week on a side hustle or learning a new skill. Always keep learning.
- How do you see Freelancing evolving in future? What qualities one should grow in the data science space to cater to client needs in 2020 say?
I think freelancing will become more common in the future as we move farther towards a globalized world. We are already seeing it with websites like Upwork and Fiverr (a place where most gigs are just $5, this is where the real price wars happen). I think the Freelance world will only become more competitive and move more towards a winner take all format. Those in roughly the top 10% of their field will probably be able to make it with those in the top 1% doing phenomenally well.
Data science will also be an involving field. First, I think the data scientist (and the worker of the 21st century) needs to constantly learn and evolve their skill set. Adaptability will be key. It’s always difficult and dangerous to predict exactly how a field will evolve. However, I believe in the next few years we will see much more movement towards cloud computing and having applications on the cloud. Enterprises will move from a server environment to a cloud environment and we’ll have access to more data than ever. The biggest problem will be sifting through a sea of information to find the useful bits than can inform decision making.
- Do you have anything to add before the end of this interview?
Pretty comprehensive interview here. I’ll just say thanks for reading this far and I am happy to chat a little for those who have other questions or follow ups. If you want to get in touch with please call 408-913-3200 or email firstname.lastname@example.org
Kalyan Banga200 Posts
I am Kalyan Banga, a Post Graduate in Business Analytics from Indian Institute of Management (IIM) Calcutta, a premier management institute, ranked best B-School in Asia in FT Masters management global rankings. I have spent 6 years in field of Analytics.