For Exercise 4.3.4 I am getting the same output with I use standard deviation or r-squared both are equivalent to the "Expected Output", but neither result allow me to get a Correct answer from the auto-grader. Since this course was originally created a while ago, I was wonder if @Abrita Chakravarty or someone from the community could verify whether something has changed or if I am just misunderstanding and getting it wrong.

Thank you again.

Hi,

I hope you could tell me what is wrong with my answer. This has bothered me so much that I joined this community.

Exercise 4.1.3

c = Classify[trainingSamplesMNIST, Method -> "NearestNeighbors"];
ClassifierMeasurements[c, testSamplesMNIST, "Accuracy"]

Output= 0.74 ---> this is also the expected output.

Exercise 3.3.2

Why is this not accepted as the correct answer:

BioSequence["DNA", 
   EntityValue[
    Entity["Gene", {"BRCA1", {"Species" -> "HomoSapiens"}}], 
    "ReferenceSequence"]] // LetterCounts[#, 3] & // KeySort

Thanks Abrita,

That clarified the intent. Autograder confirms.

Roger

Hello group,

Could someone please clarify the instructions to exercise 4.3.2. The instructions seem straightforward and I interpret them to mean that you would like all words within the given 100 word list that have equal length to be clustered or grouped into a separate sub-collection for each word length. In other words, all strings containing 13 characters are collected into one sublist, all strings containing 12 characters are collected into a different sublist, and so on. The expected output is not represented this way. Perhaps I am not understanding the instructions correctly. Please advise.

Thanks,

Roger

The built-in classifier "ProgrammingLanguage" of ex 4.1.2

runs on the cloud but not in my version 12.2 .0 desktop in Windows 10. It is not in the Wolfram Documentation for my desktop version.

I am assuming its experimental, if it is , can you just check that if it is used in a question that it is in the latest distributions of Mathematica ...... I spent an age trying to get this to run on my desktop, before in desperation running it in the cloud .

Hi @doug Beveridge, The "Programminglanguage" classifier is definitely available in the latest version of the Wolfram Language. It has been available since its introduction in 2018 with version 11.3. If you are unable to access it or use it in your version of the Wolfram Language, could you please contact the technical support team at https://www.wolfram.com/support/contact/email/?topic=technical. I am worried if it might be indicative of some other issue with your installation.

Thanks

Which documentation page are you taken to when you paste the following in the search box at the top of the documentation window? (or do you find the page does not exist on your system?)

ref/classifier/ProgrammingLanguage

Also are you able to evaluate and see the putput of the following?

Information[Classify["ProgrammingLanguage"], "Classes"]

If you do a search on the "search for all pages containing "ProgrammingLanguage" then you come to see its overloaded .

Exercise 4.3.3

After a number of trials: I think the sonnet CVIII is not in the correct cluster as demonstrated in the expected output. Please check.

Hello, Would it be possible to obtain the file File["ExampleData/StocksTemplate.nb"] referred to in session 14 Automated Report generation under the heading Automated report generation? I'd like to see it to help my understanding of the driver linkages to the template. Thank-you, John Burgers

Hi Everyone,

Thanks for making the MPDS study group a success with your active participation. We appreciate all the great questions during the sessions, the posts on this community thread, and the reports emailed to wolfram-u@wolfram.com about the issues related to the exercises. Just wanted to share some links as we wrap up the study group series:

Certificate of Completion

Requirements: Attendance at study group sessions (or watching recordings) and completing the quizzes by May 7

• https://www.wolfram.com/wolfram-u/multiparadigm-data-science/quiz-1.html
• https://www.wolfram.com/wolfram-u/multiparadigm-data-science/quiz-2.html
• https://www.wolfram.com/wolfram-u/multiparadigm-data-science/quiz-3.html
• https://www.wolfram.com/wolfram-u/multiparadigm-data-science/quiz-4.html

Level I in Multiparadigm Data Science

Requirements: Attendance at study group sessions (or watching recordings), completing the quizzes and exercises by May 14.

We understand some you are running into issues with the autograder not accepting your solution that produces the same output as shown in "Expected Output". In those cases, we are requesting you mail us at wolfram-u@wolfram.com (a description of the issue and your exercises notebook - available for download both from the course page and the download link shared during study group).

Level II in Multiparadigm Data Science

Detailed instructions here: https://www.wolfram.com/wolfram-u/certification/level2/multiparadigm-data-science/

Finally, if you are interested, more details about the Wolfram Data Science Boot Camp 2021 can be found here: https://www.wolfram.com/wolfram-u/special-event/data-science-boot-camp/

Best, Abrita

Hi Andreas, Your Level 1 certificate from the course already satisfies the requirement. Please feel free to use the discount.

Hi Vincent,

Thank you for checking in with us. We are currently in the process of manually grading the exercises submissions we have received so far (both via the online submissions on the course and via email). Thank you for your patience, as we finish up the grading. We will reach out to folks over email if we need to clarify any issues with their submissions.

best, Abrita

1. I am not able to get to "the page". How to get the page where the left side has lecture topics and " Quiz", "track the Progress"?
2. I was able to access that page until this morning but now it is not available. Did you make some changes staring this afternoon?
3. Hence, I am not able to follow your suggestion to check. 4.On the other hand, I clicked the Mathematica file and I find "open Cloud". There I am able to finish Quiz 3 to get result.
4. I have difficulty with "Exercise": It goes through the cycle of "Check Solution" and "Try Again" even with a simple coding and the correct output. It happens with almost all Exercises. There must be something, which I can't figure it out.

I just completed Quiz 3 and click "GET RESULT". But nothing happens. I tried clicking a few more times, it still does not yield anything.

How long will we be able to re-watch the web presentations?

That is really great! I expected them to self destruct, like that did in Mission Impossible. That's what happened after the Neural Network Bootcamp, which was unfortunate. There was much more information, which also seemed more complicated.

I would just like to thank the Wolfram U team for a fascinating and well presented course. Abrita, Arben and Sampurna were excellent instructors, very clear and very helpful in answering questions. I learned a huge amount and it has inspired me to go on to play with all sorts of ideas and possibilities that were raised in the course. Thank you again.

I'll just add "Hear! hear!" to that sentiment. I really enjoyed the course -- my only regret is that the course occurred at a time of year when I really would rather be outdoors, but that's on me.

Ron Goetz

Hi, Jurgen, That is the result I got, roughly 26 new words, space also counts as one word:

Hi Abrita, Arben,

I think something is wrong with your expected output for Exercise 4.5.5. When I take the output and decode it, the input sequence is as follows "GGTCTCCCAG". The original input sequence "GGCTCTTTAG" creates a different unit vector list compared to your expected output.

Am I right?

Hi Abrita, Arben,

Exercise 5.1.1 requires recasting the Dataset Association formats ...

Could either of you explain to me why this pattern matching statement doesn't operate on each element as it maps through all the elements of the association ?

easternEurope /.  KeyValuePattern[a_ -> b_] -> {QuantityMagnitude[b] -> a}

yields ...

{{6.30805*10^10, 9452409} -> Entity["Country", "Belarus"]}

Thanks, John Burgers.

Hi again John,

It seems like this is a known issue—coincidentally, somebody has commented on it internally on this very day! In the meantime, here is an alternative construction that I think should generate the output that you want:

Association@KeyValueMap[QuantityMagnitude[#2] -> #1 &, easternEurope]

Hi John—an update re: KeyValuePattern, which was the actual cause of the strange-seeming output that you saw. Basically, it works on a sequence of rules inside a List or Association, so once it found one match in your Association (in this case, for the first country, Belarus), it does the replacement on that and calls it a day. More technically, it's basically doing this:

ReplaceAll[easternEurope, 
Association[___, a_ -> b_, ___] :> {QuantityMagnitude[b] -> a}]

In any case, KeyValueMap is the more natural function to use here, as you are trying to Map some function which takes Keys and Values as its explicit inputs.

Thank-you Rohit, that's an easily readable way of doing it. John

You are right. There is an issue where ListPlot is unable to handle the multiple datasets in the EntityAssociation format. However the following should work:

ListPlot[{Values@easternEurope, Values@westernEurope}]

Now what remains, is to set up the labels for each data point. For example, for the first dataset the labels are available as Keys@easternEurope.

You will see on the documentation for ListPlot that one way to provide labels for data points is

ListPlot[{data1, data2, ...}]

where datai can have the form:

{y1,y2, ...}->{"lbl1","lbl2",...}

as well as

 {{x1,y1},{x2,y2}, ...}->{"lbl1","lbl2",...}

Daily Study Group Session Cancelled (Apr 27, 2021)

Our sincere apologies. It seems like BigMarker, our webinar platform provider, is facing some issues today. We are therefore unable to host the study group session.

We were planning to talk about Neural Networks in today's session. Instead we are asking you to look at the video at https://www.wolfram.com/wolfram-u/multiparadigm-data-science/neural-networks.html today. We'll discuss the code from the notebook at tomorrow's session.

It is a precedence issue, need parenthesis

DistanceFunction -> (Abs[StringLength@#1 - StringLength@#2] &)

Good call, Rohit—when using Rule (->) with pure functions, it's good practice to group your pure function within parentheses. Otherwise, the pure function will often be interpreted as being (so to speak) (DistanceFunction->Abs[StringLength[#1]-StringLength[#2])&, which isn't quite the right grouping.

Hi Lara!

There's no real hidden meaning here; different functions with different intents will take inputs in different orders, even if they're of the same input type (here, List and Function or Function and List). The language design is such that we'd like to have the order be intuitive—in your case, we'd like to map f onto a list, so it's Map[f,list] or f/@list, but we'd like to Group a list By some function f, so there it's GroupBy[list,f].

We don't have different classes for our functions, and as you've seen many functions can take one argument, two arguments, or tons of different optional inputs—for each individual function, we've tried to make the order of the inputs make sense for that specific function.

This, I don't have a particularly good justification for. I suspect it might be mirroring RandomChoice, RandomSample, and other such functions, where weights are specified first when given in that format (List of Rules), but I'm not sure...

Addressing some of the issues posted above:

Quiz

• Quiz 2 Problem 5 - You will find the answer on the documentation page for WordCloud http://reference.wolfram.com/language/ref/WordCloud.html
• Quiz 3 Problem 8 - The question should be rephrased "What value for AcceptanceThreshold t will return a shorter list of items for the following expression?"

Exercise

• 2.2.3 The hint is steering the solution towards use of GroupBy, but you will need to use ToLowerCase and StringTake before you can use GroupBy. And Length after you have made the groups. Note: Of course this can be done much more simply by using LetterCounts. Even if the autograder does not accept the latter solution, we will accept it if you send it in as part of your final solutions.

• 2.3.1 For example, the unique values in this list {1,5,1,1,5,5,5,1,3,1} can be found with DeleteDuplicates. Length can be used to count the number of those unique values. And you can do this for each column a data set. Transpose can be used to change columns to rows so you can easily Map your function on them. Note: Union would do the same work as DeleteDuplicates and again if the autograder does not accept your alternate solution, please email it.

• 3.3.5 Expected output should be {23.6344,35.9256,66.2097,82.9443,90.5781}

Hi Doug—I don't think that there's anything contradictory here, though it is a bit counterintuitive. The documentation can be easy to misread here, for whatever reason, so I'd highly recommend giving it another look!

Because of the context of this being a quiz, I cannot give you a specific example. I can confirm, however, that that is indeed a valid syntax and that there is no error with the question as posed.

I catch a bug in the Ex3.3.5 today:

The first column should change to the first row here, check with following:

0.338xxx is min of the first row,

Hi @Jürgen Kanz, We do appreciated your participation and feedback, as we try to resolve the various issues with the autograding of exercises. The grader has been set up to accept the variety of computations possible with the Wolfram Language and attempts to check the solutions in multiple different ways, based on the type of problem. We do use the approach you have suggested, in certain cases. We feel good about what the grader is doing right now but also realize we need to resolve a few more issues. All the feedback we are receiving from our study-group participants is definitely helping us out. Thank you.

Exercise 2.2.3

GroupBy[WordList[], StringTake[ToLowerCase[#], 1]&, Length]

Provided the answer but failed the grader. Did you have another answer in mind?

Exercise 3.3.5

We are asked to calculate some statistical figures "for the first column of the following dataset". The expected solution provides the results for the first row of the dataset. What do you really want?

Try

https://www.bigmarker.com/wolfram-u/dsg-mpds-week-1-recap?bmid=3a02139cedcf

why does this code generate an error?

Attachments:

Untitled-1.nb

You should simply determine how many different numbers occur in each column.

Since a number of study group participants are working on the exercises, we are requesting that you do not post the solutions here on community. You are of course welcome to discuss various functions and provide examples of code that do not give away the entire solution.

If you are running into issues, where the autograder is not accepting your solution (although the output is identical to what is shown as the "Expected Output"), please feel free to email us at wolfram-u@wolfram.com.

Sorry about that. The question should simply read:

Represent the information from the previous problem:

    EntityValue[EntityClass["Country", "SouthAmerica"], "GDP", "EntityAssociation"] 

in a PieChart with automatic ChartLegends. Set the ChartStyle as "DarkBands".

In Exercise 3.1.3 Visual Exploration

I am struggling to get the

to work in

getting the folowing error any ideas

Hello, Let me try to address a number of questions here.

1. Quizzes: We are addressing some issues we found with the quiz. @JuergenKanz The issue with Quiz 1 Problem 9 will be resolved in the fix we are deploying.
2. Autograder not accepting solutions to exercise problems: We will discuss this in more details during Friday's session. @UlfSchmidt and @DougBeveridge We are also in the process of setting up a way for you to download the exercise notebooks, work on them on your desktop and submit for manual grading (especially when the autograder does not accept your solution). There are so many different ways to implement a solution in the Wolfram Language... although we have set up the autograder to accept multiple solutions, we do run into issues sometimes.
3. "Download notebook" links: We are looking into the issue of the "Download notebook" not working. @DougBeveridge For now please download the zipped notebooks from the link shared during the webinar sessions.

Thank you for your patience as we try to resolve the issues. We appreciate your interest and participation in the study group sessions.

Jürgen

In general , if you use DimensionReduce without specifying a dimension it will reduce to a certain dimension (columns) . Try it out using Random Matrix's of random sizes and it will reduce to a certain dimensionality (columns) every time.

Correct. We are in the process of updating the answer options.

I agree, to edit the excercises is a nightmare. I lost the button to check the solutions for all sections.

Windows 10

Thanks , running the exercises in the cloud now

Exercise 2.3.1 / 2.3.3 / 2.3.4 I get exactly the same solution as the expected solution. The solution check answers always with "Try again". What i am doing wrong?

Abrita, I can not find the "Exercises" in The MPDS notebook. I really appreciate it if you would help.

Once you go to a new subsection of the course, the Exercise-File is copied into your private Wolfram Cloud. It is much easier to work with the copy. You can work with the files either on the cloud or locally.

No problem. We can add the "Build a Project Workflow" notebooks to the same download location.

Yes, we will share a link to download all the notebooks at once, in the study group session today. We'll also look into the issue of the "Download Notebook" links not working as expected. Thanks.

Clicking the Download Notebook link did not result in a download for the sessions {Explore, Analyze, Handling different types of data, restructuring Data}.

Is there something else required to get the downloads?

Thank-you.