Daily Time-Series Forecast Model
Budget: ₹600 – ₹601 INR
Predict daily theatre audience counts across multiple locations
In this competition, you are provided with a time-series forecasting problem centered around cinema audience attendance. The data comes from two separate booking platforms:
BookNow: an online booking and aggregation platform where users can search theatres and book tickets in advance.
CinePOS: a point-of-sale (POS) system installed at theaters that tracks on-site ticket sales.
Files
cinePOS_theaters.csv – CinePOS theatre info
booknow_theaters.csv – BookNow theatre info
movie_theater_id_relation.csv – Mapping between BookNow and CinePOS theatres.
cinePOS_booking.csv – CinePOS bookings
booknow_booking.csv – BookNow bookings
booknow_visits.csv – Daily audience counts
date_info.csv – Calendar info
sample_submission.csv – Submission format (ID = book_theater_id + show_date, audience_count).
Notes
Audience counts can be influenced by holidays, weekends, theatre type, and booking trends.
The dataset is anonymized; latitude and longitude are approximate values.
Use both POS and online booking patterns to forecast future theatre audiences.
Some theatres may be closed on certain days, leading to zero audiences. Such days are included in the dataset but are ignored in final scoring.
and you have to use only required package to do
Libraries allowed
Only the following libraries will be allowed to be used in the competition:
NumPy
Pandas
Matplotlib
Scikit-learn
XGBoost
Seaborn
Imblearn
SciPy
Pickle
regex
Lightgbm
Plotly
Please note these libraries are already installed in your Kaggle notebook. Python in-built libraries can also be used. Using libraries like TensorFlow, PyTorch, NLTK, word2vec, textblob etc is not allowed.
In this competition, you are provided with a time-series forecasting problem centered around cinema audience attendance. The data comes from two separate booking platforms:
BookNow: an online booking and aggregation platform where users can search theatres and book tickets in advance.
CinePOS: a point-of-sale (POS) system installed at theaters that tracks on-site ticket sales.
Files
cinePOS_theaters.csv – CinePOS theatre info
booknow_theaters.csv – BookNow theatre info
movie_theater_id_relation.csv – Mapping between BookNow and CinePOS theatres.
cinePOS_booking.csv – CinePOS bookings
booknow_booking.csv – BookNow bookings
booknow_visits.csv – Daily audience counts
date_info.csv – Calendar info
sample_submission.csv – Submission format (ID = book_theater_id + show_date, audience_count).
Notes
Audience counts can be influenced by holidays, weekends, theatre type, and booking trends.
The dataset is anonymized; latitude and longitude are approximate values.
Use both POS and online booking patterns to forecast future theatre audiences.
Some theatres may be closed on certain days, leading to zero audiences. Such days are included in the dataset but are ignored in final scoring.
and you have to use only required package to do
Libraries allowed
Only the following libraries will be allowed to be used in the competition:
NumPy
Pandas
Matplotlib
Scikit-learn
XGBoost
Seaborn
Imblearn
SciPy
Pickle
regex
Lightgbm
Plotly
Please note these libraries are already installed in your Kaggle notebook. Python in-built libraries can also be used. Using libraries like TensorFlow, PyTorch, NLTK, word2vec, textblob etc is not allowed.