# Datacamp Cources Rating — Data

来源：https://www.kaggle.com/competitions/datacamp-cources-rating/data

采集时间：2026-10-20T11:03:40Z

# Competition Description

Online learning platforms like DataCamp host thousands of courses on Python, R, SQL, data science, and machine learning. Each course is described by metadata such as its title, difficulty level, programming language, number of subscribers, and more.

Your challenge is to predict the **average rating of a course** (`course_rating`, a float between 0.0 and 5.0) using its metadata.

This simulates a real-world scenario for e-learning platforms: can we forecast course quality before release and identify what makes a course successful?

---

# Dataset Description

The dataset is provided in three files for this competition:

**`train.csv`**  – Training set with full metadata and the target variable **`course_rating`**.
**`test.csv`** – Test set with the same metadata but without **`course_rating`**; participants must predict this.
**`sample_submission.csv`** – Template showing the required submission format.

The dataset contains metadata for DataCamp courses, including:

- `id` (integer) → Unique identifier for each course
- `title` (string) → Name of the course    
- `programming_language` (string) → Main language used (Python, R, SQL, etc.)  
- `difficulty_level` (categorical: Beginner, Intermediate, Advanced) → Target audience level  
- `num_chapters` (integer) → Number of chapters  
- `num_exercises` (integer) → Number of exercises  
- `num_videos` (integer) → Number of videos  
- `nb_of_subscriptions` (integer) → Number of enrolled learners  
- `xp` (integer) → XP points required to complete the course  
- `time_needed_in_hours` (integer) → Approximate completion time (hours)  
- `instructors_names` (string) → Course instructor(s)  
- `course_rating` (float, target) → Synthetic average learner rating (0.0–5.0)

---

# Task

- **Target variable:** `course_rating`  
- **Type:** Regression  
- **Goal:** Build a model that predicts the course rating from its metadata using train.csv and submit predictions for test.csv.

---

# Evaluation

Submissions are evaluated using **Root Mean Squared Error (RMSE)

---

# Inspiration & Use Cases

- Which features (difficulty, programming language, number of subscribers) are most predictive of high ratings?  
- Do “Beginner” courses receive better ratings than “Advanced” ones?  
- Can text analysis on course titles and descriptions improve predictions?  
- How do instructor reputation and course length affect ratings?

---