What's it like being a data scientist?
According to an article on Indeed, data scientists "may work in social media companies and use the user data to understand habits and suggest content that the users will enjoy. They may also work in politics and make predictions about the election results. In general, data scientists use the data from their analyses to solve real-world problems in business and politics." What other fields do data scientists use their knowledge and skills?
Data Scientists are also incredibly useful in today's policing, most especially in large metropolitan departments. A law enforcement agency's CAD (CPU-Aided Dispatch), E-Cite citations reports, incident reporting system, E-Crash accident reporting system, phone logs and more all contain useful but sometimes unharvested data on incident and call-for-service times, GPS locations &/or specific neighborhoods/zones, felony vs misdemeanor arrests, category of offense (traffic or criminal), number of officers needed to contain or assist with the incident or call for service, etc.
Maybe you're asking, "How can such data be used?" A competent data scientist could compile all this data into a more tangible respresentation such as a spreadsheet to help a Police Chief make educated decisions on how to best position his or her officers and schedule his or her dispatchers to reduce response time, deter crime, prevent traffic accidents, and so on.
I believe that-with the growing critical shortages of policemen and women across the nation-data scientists are likely the new future of Law Enforcement. It is in this way that most of the guesswork is removed and with it, the inefficiencies built into the old guesswork system.
I hope this helps to add to your list of options to explore, hopefully via shadowing a Data Scientist in some of the many the different applicable professions. Keep inquiring and exploring!
Cheri recommends the following next steps:
Let me first explain about the job. "Data Scientist" is applied to a whole range of jobs that generally involve some mixture of mathematics, statistics, computer programming and domain knowledge (knowledge of banking, or knowledge of social media, etc.). The precise mix of those ingredients vary greatly. For example, it's common in tech companies to have a job "Applied Data Scientist" that is very heavy on computer programming, whereas "Research Data Scientist" is less heavy on computer programming but heavier on the statistics and math.
The industry makes a big difference too. I started my career in banking, but moved to healthcare and eventually tech. Healthcare and tech are probably the most starkly opposite and for good reason. Data scientists' recommendations in healthcare can have significant impact on people's lives (and deaths). Consequently there is a very heavy regulatory burden, and you can expect to work on projects for a very long time until you have turned every stone and are absolutely sure your recommendation is right. Tech is very different. Take streaming for example. For most data scientist recommendations in streaming, the worst thing that can happen is some folks can't stream for 15 minutes or however long it takes someone to revert back the code. In other words, the cost of a mistake is (relatively) cheap. So you can expect a rapid fire succession of projects which can be exhilarating for some, or dizzying for others.
In my opinion, the biggest difference is the culture of the company where you work. This is what will dictate whether you work 30 hours per week, 40 hours per week or 80 hours per week. Most companies don't have time hour expectations written down, but if you join a company where everyone is working 80 hours per week, it's going to be very hard for you not to do the same. Company culture also dictates how cooperative or competitive your co-workers will be. There will always be a mix of cooperation and competition, but to which degree depends on the company. Having an impartial insider to talk to is invaluable before joining a company. Finally, the company will play a huge role in how and who you interact with. For example, if you join a startup you will wear many hats including (most likely) talking directly to the person whose problem you are trying to solve. With larger companies, you will generally be more removed from the actual problems. Some folks find the separation comforting - some find it frustrating.
With regards to real world applications, here are just some of the things I've worked on over 25 years in data science:
- Predict whether a consumer would pay back a loan or not
- Predict a bank failure based on the predictions above
- Identify the best targets for a marketing campaign
- Speed up an operational process while improving its quality
- Identify patients in need of help to stay on their medications
- Identify patients who need extra help when starting Opiods
- Quantify the impact of the "help" provided to patients
- Predict how many people will see a movie that's never been shown before and for how long
- Make recommendations to improve the quality of a streamer's catalog without increasing the cost
In short, data science can be applied to an enormous range of problems. I would say, anything that can be measured can be a subject of data science. Which brings me to my last point. If you enjoy computer programming and statistics/mathematics, then data science is almost certainly a good match for you. I wouldn't worry about the industry, company or job before college. I think you will find that there is such a wide array of these that you will be able to find a job you love. Before college, focus on learning the fundamentals of computer programming and statistics/mathematics. When you're ready to decide on your college major, then think about industries and consider doing a double major which can be a real differentiator in the job market. For example, if you decide that you want to apply data science to film, then a double major in film and statistics or film and computer science will set you apart from the vast majority of applicants.
I hope you remain interested in data science. For my part, I've enjoyed every minute of my journey in data science - I hope you will too.
There is actually quite a thin line between Data Science, Machine Learning and Artificial Intelligence. and the best part is all the top notch techniques start with Data Science.
As a Data Scientist, we will often find us discussing different Processes with customers, internal or external. We will find that people have business requirements from the data, Examples like explained in the other answers here. These requirements are sometimes concrete, but most of the time, they would have abstract requirements.
According to me, the best part of being a Data Scientist is this piece, when we are trying to discover what story the data is telling us and how does it match with the requirements. Do we need something additional there, or does the data in hand suffice.
First of all, don't be surprised if we are spending lot of time, in cleaning up the data, making certain assumptions, validating those assumptions with our customers before even we start making a good sense on the further piece. With lot of tools available at our hand, nowadays, its not actually that big a task, but definitely takes time and effort, not maybe with the code, but with all those assumptions.
Then we start analysing and visualizing the data, the trends, the charts and the story. While there are chances, that we get carried away in the process, its always important, to go back to the customers and validate/ take their views on these initial findings.
The final piece is the Modelling and Output visualization. We will normally that its not a coding job at all. Hundreds of reusable codes will be easily available, just with a few googling, or directly using Chat GPT now. The important point to note here is we have to be completely conversant with the statistics behind those models, the parameters which need tweeking to fine tune the model.
Then finally presenting the Model Outputs to customers. Make sure you are using models which you are able to explain to them, because nobody is going to just take any conclusions blindly or just because some xyz model is giving the output.
For me, these interactions with customers, understanding processes, the discovery phase, analysis phase make this an awesome job. Coding is actually easy and simple...the fun part is applying our own intelligence.
I haven't put any tools or roadmap here, because I am sure you can easily find it with a bit a googling and help from ChatGPT.
I hope you continue your path on Data Science and explore new opportunities in this field.
The data scientist will try to co-relate things and see whether it is making sense for the business. You need to be curious and ask many questions about the business from various levels.
Some of the simple questions may solve bigger problems in the business. Try to learn some tools which will make your job (crunching the data) easier but all you need to be curious to solve the problem or wanting to know the root cause.