Data Science/ML Problem Creation.

For Client: Go to the library and select my company questions. Then click the + button at the bottom right to create questions.

For Freelancer/Creator: Go to the account and select avatar on the top right and click the Content Dashboard. 

• Then click the + button at the bottom right to create questions. 
• In the Create New problem dialogue box, fill in the problem name, choose the level of difficulty, and select Problem Type as Data Science. 

• In the Problem Statement page, specify:
• Problem Name: The problem name should be short and topical. It should not give any hints to solve the problem.

• Expected solving time (in minutes): The time required to solve the problem, should be an integer between 1 to 999.
• Description: A good problem statement should comprise a clear and detailed problem description, at least one sample output.
• Difficulty level: Set the difficulty level for your Question by associating the tags as Easy, Medium, or Hard. 
• Scoring: You can leave it default or change it accordingly with a penalty mark for the wrong submission.

• Maximum re-submission allowed: We can restrict the number of submissions of the solution, leaving it 0 for no limit.
• Execution time limit: It's to stop the execution of the code after a given time, leave it blank to use default limits.
• Allowed programming languages: Select the allowed languages to answer this particular problem or blank empty for allow all languages.
• Discovery tags, Insight tags: Tags are words or phrases that help with the searchability and organization of your questions. You can add the existing tags or create new tags.
• Primary Technology: Select a primary technology that best represents your problem.

• Datasets: Upload datasets that are required for training models, and for evaluation. These datasets should be in .zip files up to 20 MBin size each.
  • Training Dataset: Dataset that candidates use to train their models. When the training dataset .zip file is uploaded, it unzips the file and stores it in the location “ /data/training/ ”. For example, if the training dataset .zip file has file1.csv, file2.csv, file3.csv.Then the location to access these files is: 
    • /data/training/file1.csv
    • /data/training/file2.csv
    • /data/training/file3.csv
  • Evaluation Dataset: Dataset that candidates use to predict the outcome. The target variable will not be provided in the test dataset. An evaluation dataset is used when you submit the problem. 
    • Name: test.csv
    • Location: /data/test/test.csv
  • Validation Dataset: Format that candidates should follow to create their submission file. A validation Dataset is used when running the problem.
    • Name: sample_submission.csv
    • Location: /data/test/test.csv
• Stubs: In the Stub section, click + ADD NEW to generate the code stub and specify the technology (programming language), and optionally add necessary imports and code, then click SAVE to generate the stub.

• Sample solutions: Click on +Add New, select the technology (programming language), and add a working solution to the problem.
• Testcases:In the test cases section of the displayed page, perform the following operations:
  • Click +Add New and Specify the name of the test case.
  • Select the programming language for your test case code.
  • Write your test case code for evaluating the solution. 
  • Enter test case weightage for the test cases.  If the code passes a particular test case, the score is assigned. Test case score = (test case weight/total weight)*total score.
  • Optional: Select the Is sample test case? check the box if you want the test case to be a sample test case.
  • Click +ADD TESTCASE.

• Attachments: Add any extra file required for test cases. Use the file link for the file path.
• Click on Save Problem to save the question.