{
  "fetched_at": "2026-10-19T15:23:42Z",
  "source_url": "https://www.kaggle.com/competitions/dev-o-thon-23/data",
  "pages": [
    {
      "name": "data-description",
      "content": "**Title:** Lead Scoring Algorithm Development Dataset\n\n**Test Description** : The \"EST.csv\" is a csv file specifically designed to test your model . The Answer should be in csv format containing ID,Status Columns.\n\n**Train Description:**\nThis dataset contains information about leads that a company has acquired. The purpose of this dataset is to develop a lead scoring algorithm, which will predict the likelihood of a lead becoming a customer based on their demographic and behavioral data.\n\nThe dataset is provided as a CSV file named 'Rain.csv'. It contains **28 columns** and more than **100,000 rows** of data. Each row represents a lead, and each column contains a different attribute of that lead. The columns are as follows:\n\n**Email:** The email address of the lead\n**Leads_Source**: The source through which the lead was acquired\n**Work email**: The work email address of the sales person\n**Job title**: The job title of the sales person\n**Sales owner:** The name of the branch owner assigned to the lead\n**Owner Email**: The email address of the salesperson assigned to the lead\n**Territory**: The geographical territory of the lead\n**Team**: The team assigned to the lead\n**Product_interested**: The product(s) that the lead has shown interest in\n**Lead_Zipcode**: The zipcode of the lead's location\n**Lead_State**: The state of the lead's location\n**Lead_city**: The city of the lead's location\n**url-campaign**: The URL campaign through which the lead was acquired\n**page-url**: The webpage from which the lead was acquired\n**Lead_hour**: The time of day when the lead was acquired\n**Fulfillment Status:** The status of the lead's order fulfillment\n**Fulfilled at**: The date and time when the lead's order was fulfilled\n**Accepts Marketing**: Whether the lead has given permission to receive marketing communications\n**Shipping**: The shipping method used for the lead's order\n**Total**: The total value of the lead's order\n**Discount Amount:** The amount of discount applied to the lead's order\n****Payment Method:**** The payment method used for the lead's order\n**Refunded Amount**: The amount refunded for the lead's order\n**Vendor**: The vendor associated with the lead's order\n**Outstanding Balance**: The outstanding balance on the lead's order\n**Source**: The original source of the lead\n**Status**: The Target Value of The Lead\n**ID**: The unique identifier for the lead\n\nThe target variable is explicitly defined in the dataset and it is implied that the goal is to predict which leads are most likely to become customers. The algorithm will be evaluated on the basis of its recall and precision metrics.\n\nThis dataset is suitable for data scientists and machine learning engineers who are interested in developing lead scoring algorithms. It can also be used to study customer behavior and to identify trends in lead acquisition and conversion.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "\n# Rules\n\n## Participation\n\n- You must be a student from an Indian college with age group 18-25 years to participate in this competition. You must provide proof of your identity and enrollment status to be eligible for the prizes , if asked.\n- You may participate as an individual or as a team of up to three members. If you participate as a team, you must designate one member as the team leader who will be responsible for submitting your predictions, code, and report.\n- You may only submit one entry per team or individual. If you submit multiple entries, only the submission referred in typeform submitted will be considered. Typeform Link Will be Shared Soon.\n- You must not share your code or data with other teams or individuals, or use external data sources or pre-trained models that are not publicly available. You must acknowledge all the sources and libraries that you use in your code and report.\n- The prizes will be awarded after verifying the eligibility and validity of the submissions. The organizers reserve the right to disqualify any submission that violates the rules or the spirit of the competition.\n\n## Code of Conduct\n\n- You must respect the privacy and confidentiality of the data provided for this competition. You must not use the data for any purpose other than participating in this competition, or share the data with anyone outside of this competition.\n- You must abide by the Kaggle Code of Conduct and Terms of Use. You must not engage in any fraudulent, abusive, or unethical behavior that may harm the integrity or reputation of this competition or Kaggle.\n- You must accept the decisions of the organizers as final and binding. The organizers reserve the right to modify or cancel the competition at any time without prior notice.\n\n****\nBy participating in this competition, you agree to abide by these rules and acknowledge that you have read and understood them.\n****",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "References ",
      "content": "#### Reference & Inspiration for Participants\nThere are various prebuilt models based on **lead-scoring algorithms**. When developing a **lead-scoring model**, it can be helpful to keep in mind prebuilt lead-scoring companies that offer similar services. These companies can provide inspiration for your algorithm development and help you to understand the types of features and techniques that are commonly used in the industry. Some examples of prebuilt lead-scoring companies include:\n\n\n    **Infer**: Infer offers a lead-scoring platform that uses predictive analytics and machine learning to score leads based on their likelihood to convert. \n\n\n    **MadKudu**: Their platform provides a simple dashboard that allows sales and marketing teams to easily prioritize their leads based on their likelihood to convert.\n\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "**Introduction**\nIn **Dev-a-thon,** participants are challenged to develop a **lead-scoring algorithm** that can accurately predict the likelihood of a customer making a purchase on an e-commerce platform based on** accuracy and recall value**.\n\n**Recall value** is an essential metric in lead scoring, as it measures the algorithm's ability to correctly identify all positive cases (i.e., customers who are likely to make a purchase) out of all actual positive cases in the dataset. In other words, it measures how well the algorithm is able to capture all potential customers who are likely to make a purchase.\n\nThe competition judges the participants' algorithms based on both accuracy and recall value. The winning entry is the one that provides the **most accurate predictions while also achieving the highest recall value**. This ensures that the algorithm is accurate and effective at identifying potential customers who are likely to make a purchase. The competition is designed to provide valuable insights and solutions to** e-commerce** businesses looking to increase their sales and revenue by accurately identifying potential customers who are likely to make a purchase.\n\n**Email : [ylp@growthjockey.com](mailto:ylp@growthjockey.com) , [diviyansh.modi@growthjockey.com](mailto:diviyansh.modi@growthjockey.com)**\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Judging & Prizing",
      "content": " \n**Judging & Prizes:** The evaluation of the model will be based strictly on its **recall value and accuracy**, as previously mentioned. To conduct the evaluation, the provided **.pynb** file will be tested using the same parameters as the original Excel file, but with modified values.\n\n**Growth Jockey's YLP** has successfully organized several events, including **Cas-a-thon 2k23 and Dat-a-thon 2k23**, where participants have received over **INR 50000** in **E-Commerce Cash Vouchers**. As an acknowledgment of their efforts, participants were awarded a certificate of merit and given the opportunity to attend an exclusive founder's talk. Moreover, participants had a valuable chance to network with Growth Jockey's leadership team. The prizes of the **Dev-a-thon** Include prizes worth upto **1 Lakh Rupees**.\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Getting Started",
      "content": "\n**Getting Started**\nHere are few Suggestions for the participants so to smoothen their journey through Make-o-thon \n\n\n\n1. **Understand the problem:** The problem is to develop an algorithm that can accurately predict the likelihood of a customer making a purchase on an e-commerce platform based on both accuracy and recall value.\n2. **Gather and preprocess the data**: Growth Jockey will provide you with a dataset containing information about various customers. The dataset may contain various features such as sale date , subscribed to newsletter etc.  Preprocessing the data involves cleaning the data, removing any irrelevant features, and normalizing the data to make it more suitable for analysis.\n3. **Develop and test your model**: Once you have preprocessed the data, you can start developing your lead-scoring algorithm. You can use various machine learning techniques such as logistic regression, decision trees, or random forests to develop your model. Once you have developed your model, you should test it using cross-validation techniques to evaluate its accuracy and recall value.\n4. **Submit your entry** After testing your model, you can submit your entry on the platform. The entry will be a .pynb file which further will be tested on different dataset values with the same parameters. They will evaluate your entry based on both accuracy and recall value, and the entry with the best overall performance will be declared the winner.\n\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Score",
      "content": "\nThe goal of the Dev-o-thon competition is to develop a lead scoring algorithm that can predict the probability of a customer converting into a buyer based on various features such as demographics, behavior, preferences, etc. The Benchmark Values for Submission in the Competition include following : \n\n| Score Matrix | Expected Score |\n| --- | --- |\n| F1 Score | 50% |\n| Precision %'age | 66% |\n| Recall %'age | 70% |\n\n\n## Submission Format\n\nThe submission file should have a header and contain two columns:\n\n- `id`: the unique identifier of each customer\n- `score`: the predicted label of conversion (0 or 1)\n\nFor example:\n| ID | Status |\n| --- | --- |\n| 985asdf5 | 0.0 |\n\n\n## Evaluation Metric\n\nThe submissions will be evaluated on the **F1 score with micro averaging**. The F1 score is the harmonic mean of precision and recall, and it ranges from 0 (worst) to 1 (best). Precision is the ratio of true positives to all predicted positives, and recall is the ratio of true positives to all actual positives.\n\nThe F1 score with micro averaging can be calculated as follows:\n\n```\nF1_micro = 2 * (TP / (TP + FP)) * (TP / (TP + FN)) / ((TP / (TP + FP)) + (TP / (TP + FN)))\n```\n\nWhere TP is the total number of true positives across all classes, FP is the total number of false positives across all classes, and FN is the total number of false negatives across all classes.\n\nThe F1 score with micro averaging gives equal weight to each class and each sample, and it is suitable for imbalanced datasets.\n\nThe higher the F1 score, the better the performance of the lead scoring algorithm.\n\n## Leaderboard\n\nThe leaderboard will show the public F1 score of each submission based on a subset of the test data. The final ranking will be determined by the private F1 score based on the remaining test data, which will be revealed after the competition ends.\n\n",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Timeline",
      "content": "- The Competition officially begins at IST April 14 , 2023 5PM & ends at April 16,2023 11.59PM IST\n- Participation done in the due timeline will only be considered.\n- Participants must submit their final typeform before April 17 , 2023 5PM IST. Please make sure that in case of duplicate typeform responses , the one with higher recall value will be choosed.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}