#189 overall
Data Analysis and Statistical Inference
Coursera
2 recommenders · 2 threads · recommended in 2 years, 2015–2016 · score 2.4
An introductory Coursera course from Duke University covering data analysis and statistical inference. It is the statistics course commenters point to as a stronger alternative to the statistics module of Coursera's data science specialization.
Commenters call it excellent and recommend it to fill the statistics gap left by the Coursera data science specialization.
Mathematicsbeginnerfreemiumvideo lectures#statistics#statistical inference#data analysis#duke#coursera
Recommendations by year
20062026
I also completed this specialization, but unpaid so did not do the capstone project. I agree with most of your take/skip assessments. The Statistical Inference course was particularly disappointing. I recommend the excellent Data Analysis and Statistical Inference (https://www.coursera.org/course/statistics) to fill in that missing area of learning.
Coursera has a 'specilization' on Data Science. It's good, not great though. It costs about US$500.
It's not really a boot camp, but a MooC. There are 9 courses and first 3 courses are really easy. For a beginner they might be intimidating, so it depends on your prior experience. The course #6 is on Statistics which is really bad. The classes are not that explanatory and exercises seem very difficult. So, I suggest you to do another Coursera course on statistics by Duke university. This course is usually referred as DASI.
Currently I am at course #6. I have to complete 3 more and a capstone project at the end to complete the data science specialisation
I've been following the MOOC wave since the "early 3" Stanford classes back in Fall 2011. I've continued both on Udacity and Coursera partly auditing and partly participating.
The fiasco with the MOOC class, IMHO, was probably avoidable.
For one MOOCs of the Coursera/Udacity scale have been around for little over a year and I doubt even Andrew Ng/Daphne Koller or Sebastian Thrun would call themselves experts in what a MOOC is. No one knows yet and we are in a transient phase while this new knowledge works its way though the academic environment. What MOOCs will be next year is very different from what they are today. So for someone other than a founder of a MOOC to teach a class on MOOCs raises some questions. Creating models of what a MOOC is while the area is in such flux only creates more noise in the channel. For another how is it a surprise that 40K+ people signed up, for any MOOC instructor at Coursera, let alone for a class on MOOCs.
It's unclear how Coursera let a class on this topic be offered not even counting the QA issues.
While OP raises many issues in the blog post about this class and the instructor there are similar patterns in many other classes.
MOOCs as they exist today are massively instructor/university/vendor centric. This is partly because they are an incremental step away from a university experience and instructors are leraning how to play in this new arena. Also because they are free and the value to the student just of the content and structure is huge. So the motivation to expend the effort ot make it additionally student centric is small. The current state of the art is more than good enough and will be for some time.
Even then there are somethings I have seen on Coursera which could be relatively easy to do and improve the experience for everyone.
There needs to be some sort of experience sharing forum for course instructors so knowledge sharing takes place. This seems to be missing as evidenced by comments in the first week by many instructors asking for patience as they work out the kinks and learn the ropes.
Specifically instructors need to have data on previous instructors' class experience - what worked what didn't. Maybe this is available but perhaps it's not being used enough.
Suggestions to instructors from what I see over the last year
* Think through your class requirements will it scale - this is not your university class on video - it's a massive beast that needs to be fed.
* Will it be secure on an Internet scale. For instance the absolutely excellent Data Analysis class by Jeff Leek of Johns Hopkins. He had to remove a requirement for people to submit executable R code in assignments bec it would run on other people's computers during peer-evaluation. Not criticizing this class but asking will future classes have this data point available to them? Not sure.
* any special things you need to do for your class (such as the Google spreadsheet usage in the class mentioned by OP) - has it been done before at the scale. Do you have TA's or support staff that have done this before? Pl. don't use the class as a scaling experiment.
* be available on the forums if possible or make it some TA's job to monitor all forums (not sure how one person could do this especially as the class progresses) Jeff Leek being live on the forums makes the class extra special.
* if possible create weekly video "screenside chats" like Jennifer Widom did in the very first online Intro to Database class from Stanford - they were excellent as were the Google Hangouts by Thrun/Norvig in the AI class. Andrew Ng's closing statement after his Machine Learning class was very deeply felt. While student centricness is not expected at this stage - communicating with students in this fashion makes your class much more engaging and you much more human. Also it's likely (wild ass guess) that completion rates might improve with this - worth investigating. I also remember emails from the AI class exhorting students not to give up. In contrast I have received emails from classes (not to be named) congratulating me on how awesomely I have done when I haven't even logged in once :-)
* as far as # of quiz attempts and using late days there is a wide spread of parameters across classes. My suggestion as a working professional is that 5 late days for say 8 quizzes and 8 assignments is not enough. In addition I found that a 4% penalty per late day was probably much more reasonable than a 50% penalty. Other than that 1 quiz attempt is too little but 100 is too many somewhere in the region of 5 seems more appropriate.
That's all for now - hope this is helpful.
Well...first off, Big Data isn't in IT. Replacing hard drives, imaging servers, and configuring IIS/LAMP is IT.
Depending on where you are in your life and what options you have available, you could go back and get another BS, in computer science or similar. Or you could get a professional MS in computer science. Or maybe just take some coursera courses like this one: https://www.coursera.org/course/dataanalysis
Since you already have a JD, I'd think you could combine your legal knowledge with coding knowledge, possibly working in the legal department of a medium sized software company, as a bridge between legal and dev?