9
$\begingroup$

I am a 35 year old IT professional who is purely technical. I am good at programming, learning new technologies, understanding them and implementing. I did not like mathematics at school, so I didn't score well in mathematics. I am very much interested in pursuing a career in Big Data analytics. I am more interested in Analytics rather than Big Data technologies (Hadoop etc.), though I do not dislike it. However, when I look around in the internet, I see that, people who are good in analytics (Data Scientists) are mainly Mathematics graduates who have done their PHds and sound like intelligent creatures, who are far far ahead of me. I get scared sometimes to think whether my decision is correct, because learning advance statistics on your own is very tough and requires a of hard work and time investment.

I would like to know whether my decision is correct, or should I leave this piece of work to only intellectuals who have spend their life in studying in prestigious colleges and earned their degrees and PHDs.

$\endgroup$
7
  • 2
    $\begingroup$ I think this might be a bit broad for StackExchange, and perhaps considered off-topic if it concerns career advice, but see what others think. $\endgroup$
    – Sean Owen
    Oct 6, 2014 at 15:12
  • 7
    $\begingroup$ Don't forget that the people you are comparing yourself against are those that have the knowledge to have well-read blogs, have high stack exchange reps, etc, ie, not a representative sample. You are comparing yourself with the best, not the average. If you are a smart IT guy and you want it badly enough, it is there for the taking. Data is growing exponentially, our ability to analyse and manage it, possibly more slowly. So, there are plenty of opportunities, just grab the bull by the horns. $\endgroup$ Oct 7, 2014 at 19:36
  • 1
    $\begingroup$ Every company is different I guess, but in my company we don't do any insane statistics/mathematics. There are a lot of common sense problem solving though. I personally wish that my computer science background was stronger. I'd rank the skills in order of value like this: 1) Common sense, 2) Computer Science / Programming 3) Mathematics / Statistics. $\endgroup$
    – Akavall
    Oct 11, 2014 at 3:54
  • $\begingroup$ You might want to read my related answer. $\endgroup$ Jan 10, 2015 at 9:28
  • $\begingroup$ If you're a good programmer then you probably already use quite a bit of math. I can't imagine a programmer that is good and doesn't use math on a daily basis. What is the highest level of math you've used? What programming language do you use and what do you use it for? You certainly don't need a PhD to do data science, but the math is essential. $\endgroup$
    – Amstell
    Jan 10, 2015 at 19:51

7 Answers 7

4
$\begingroup$

Due to high demand, it is possible to start a career in data science without a formal degree. My experience is that having a degree is often a 'requirement' in job descriptions, but if the employer is desperate enough, then that won't matter. In general, it's harder to get into large corporations with formalized job application processes than smaller companies without them. "Knowing people" can get you a long way, in either case.

Regardless of your education, no matter how high demand is, you must have the skills to do the job.

You are correct in noting that advanced statistics and other mathematics are very hard to learn independently. It is a matter of how badly you want to make the career change. While some people do have 'natural talent' in mathematics, everybody does have to do the work to learn. Some may learn more quickly, but everybody has to take the time to learn.

What it comes down to is your ability to show potential employers that you have a genuine interest in the field, and that you will be able to learn quickly on the job. The more knowledge you have, the more projects you can share in a portfolio, and the more work experience under your belt, the higher level jobs that will be available to you. You may have to start in an entry level position first.

I could suggest ways to study mathematics independently, but that isn't part of your question. For now, just know that it's hard, but possible if you are determined to make a career change. Strike while the iron is hot (while demand is high).

$\endgroup$
2
  • $\begingroup$ I mentioned about my being weak in mathematics, during school days. I have started liking mathematics ever since I have seen its real use in solving real life problems :). So, you can suggest me ways to study mathematics. I like your answer. $\endgroup$
    – KurioZ7
    Oct 9, 2014 at 10:38
  • $\begingroup$ I always like to learn about the software problem I'm trying to solve, then learn the mathematics needed to solve the problem. However, it is possible that you won't be able to just pick up the new math and use it right away, depending on your skill level. Be honest with yourself and choose a software problem that has math you think you could pick up. Work on it daily, as part of your portfolio. Broaden your math knowledge with online courses if you find engaging software problems with math you don't understand. The key thing is habit - make time to study or code every day. $\endgroup$ Oct 9, 2014 at 16:59
1
$\begingroup$

You should look more into the infrastructure side of things if you don't like maths. The lower you go in the software stack, the further away you get from maths (of the data science sort). In other words, you could build the foundation that others will use to create the tools that will serve analysts. Think of companies like Cloudera, MapR, Databricks, etc. Skills that will come in handy are distributed systems and database design. You are not going to be become a data scientist without maths; that's a ridiculous notion!

$\endgroup$
1
$\begingroup$

In my experience to have a PhD doesn't mean necessarily be good in the enviroment of data science company, I work as data scientist and I'm just an engineer but I've known some universitary teachers who works in collaboration with my company and sometimes I've said them that Their point of view was not right because despite of their ideas and reasonings were right they are not applicables to the company activities, so we had to modify some data models to make them usefull for the company and the results lost their value so we had to seek new models. What I mean is that Data Science is a multidisciplinar area so many different people working together is needed so I think that your skills could be very useful in a data scientist team, you only have to find where you fit ;)

$\endgroup$
1
$\begingroup$

May be it will be a little offtopic, but I'd like to highly recommend you to go through this MOOC https://www.coursera.org/course/statistics. This is a very good and clear introduction to statistics. It give you a base principles about core field in data science. I hope it will be a good start point for beginning friendship between you and statistics.

$\endgroup$
1
$\begingroup$

I haven't seen this mentioned, but it's important to keep in mind that you may see a decrease in salary. I say this without knowing how much you make, but moving from (I assume) an experienced IT professional to an entry level data scientist level may not earn you as much.

Here's a link to the a portion of the 2015 Burtch Works study on Data Science salaries:

http://www.burtchworks.com/files/2015/05/DS-2015_Changes-in-Base-Salaries.pdf

As you can see, the median salary for level 1 individual contributors is 90k (across the nation). The full report has the breakdown based on region but again, assuming you're an experienced IT professional, you're probably making more than that.

Anecdotal story with n=1: One of my classmates in my DS masters program was an experienced Java developer with a house, family, etc. Although he was very interested in data analytics (paid for the program out of pocket) his potential salary doing data analytics wouldn't be able to support the lifestyle he currently had as a Java developer. As a result he essentially "wasted" his degree and went back to development. I would really hate to see that happen to more people.

$\endgroup$
1
  • $\begingroup$ Interesting information Jake! $\endgroup$
    – KurioZ7
    May 11, 2015 at 8:09
0
$\begingroup$

Keep in mind that "big data" is an increasingly trendy thing for a company to say they're involved in. Higher ups might read an article about it in HBR, and say to themselves, "I've got to get me some of that" (not that they're necessarily wrong).

What this means for you is that the advanced analytics isn't as necessary for that company as just getting something up and running might be.

Luckily for you, most of the components said companies might need are free. Moreover, I believe both Hortonworks and Cloudera have free "sandbox" virtual machines, which you can run on your PC, to play around with and get your bearings.

Advanced analytics on big data platforms are valuable, to be sure, but many companies need to learn to crawl before they can run.

$\endgroup$
0
$\begingroup$

This is a really strange question in my opinion. Why you're going to move in a new direction if you are not sure that you love this new direction or at least find it very interesting? If you do love Big Data, why do you care about the PhD intelligent creatures that are already in the field? The same amount of PhD creatures are in every area of IT. Please have a quick read at this very nice article http://www.forbes.com/sites/louisefron/2013/09/13/why-you-cant-find-a-job-you-love/ and then ask yourself if you love Big Data enough and you are ready to add your grain of sand to the mountain of knowledge

$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.