It is revealed that site 4 is UC Berkeley. Fortunately, the university provides its energry consumption data publicly.
\n\nThe following kernel extracts the historical energy consumption data for several buildings. Even though the dashboard shows one year data maximum, rest service returns data from 2016.
\n\nAs you might realize, some buildings have same square feet and built year values in both ASHRAE and Berkeley data set. Besides, some buildings have very close square feet area and same built year. There are 49 buildings in Berkeley correlated to ASHRAE data set.
\n\nhttps://www.kaggle.com/serengil/ucb-data-leakage-site-4
", "messages": [ { "id": "677942", "postDate": "11/20/2019 19:54:30", "content": "It is revealed that site 4 is UC Berkeley. Fortunately, the university provides its energry consumption data publicly.
\n\nThe following kernel extracts the historical energy consumption data for several buildings. Even though the dashboard shows one year data maximum, rest service returns data from 2016.
\n\nAs you might realize, some buildings have same square feet and built year values in both ASHRAE and Berkeley data set. Besides, some buildings have very close square feet area and same built year. There are 49 buildings in Berkeley correlated to ASHRAE data set.
\n\nhttps://www.kaggle.com/serengil/ucb-data-leakage-site-4
", "rawMarkdown": "It is [revealed](https://www.kaggle.com/c/ashrae-energy-prediction/discussion/115698#latest-674237) that **site 4 is UC Berkeley**. Fortunately, the university provides its energry consumption data [publicly](https://engagementdashboard.com/ucb/ucb/). \n\nThe following kernel extracts the historical energy consumption data for several buildings. Even though the dashboard shows one year data maximum, rest service returns data from 2016.\n\nAs you might realize, some buildings have same square feet and built year values in both ASHRAE and Berkeley data set. Besides, some buildings have very close square feet area and same built year. There are **49 buildings** in Berkeley correlated to ASHRAE data set.\n\n[https://www.kaggle.com/serengil/ucb-data-leakage-site-4](https://www.kaggle.com/serengil/ucb-data-leakage-site-4)", "votes": null }, { "id": "677951", "postDate": "11/20/2019 20:05:00", "content": "Potentially, I am starting to suspect that all sites might be publicly collected at this point. :( Is it intentional to screw up the nature of kaggle?\n━━━━━┓\n┓┓┓┓┓┃/ . `(와장창)\n┓┓┓┓┓ : . ＼○ノ
\n\n┓┓┓┓┓┃ ` /
\n\n┓┓┓┓┓┃ ノ) .\n┓┓┓┓┓┃ ,\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃
", "rawMarkdown": "Potentially, I am starting to suspect that all sites might be publicly collected at this point. :( Is it intentional to screw up the nature of kaggle?\n━━━━━┓\n┓┓┓┓┓┃/ . `(와장창)\n┓┓┓┓┓ : . ＼○ノ\n\n┓┓┓┓┓┃ ` /\n\n┓┓┓┓┓┃ ノ) .\n┓┓┓┓┓┃ ,\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃", "votes": null }, { "id": "680423", "postDate": "11/24/2019 16:31:27", "content": "Initial kernel I've published found 49 buildings but currently 74 buildings exactly matched in both UC Berkeley and ASHRAE data sets. This is equal to 1.286.941 instances in submission. Notice that submission set was 41.697.600 instances. This means that site 4 leak data is 3.08% of all submission data.
", "rawMarkdown": "Initial kernel I've published found 49 buildings but currently 74 buildings exactly matched in both UC Berkeley and ASHRAE data sets. This is equal to **1.286.941** instances in submission. Notice that submission set was **41.697.600** instances. This means that site 4 leak data is **3.08%** of all submission data.", "votes": null } ], "comments": [ { "id": 677951, "author_name": "hanjoonchoe", "author_url": "", "post_date": "11/20/2019 20:05:00", "content": "Potentially, I am starting to suspect that all sites might be publicly collected at this point. :( Is it intentional to screw up the nature of kaggle?\n━━━━━┓\n┓┓┓┓┓┃/ . `(와장창)\n┓┓┓┓┓ : . ＼○ノ
\n\n┓┓┓┓┓┃ ` /
\n\n┓┓┓┓┓┃ ノ) .\n┓┓┓┓┓┃ ,\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃
", "votes": null, "replies": [] }, { "id": 680423, "author_name": "serengil", "author_url": "", "post_date": "11/24/2019 16:31:27", "content": "Initial kernel I've published found 49 buildings but currently 74 buildings exactly matched in both UC Berkeley and ASHRAE data sets. This is equal to 1.286.941 instances in submission. Notice that submission set was 41.697.600 instances. This means that site 4 leak data is 3.08% of all submission data.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "677942": "It is [revealed](https://www.kaggle.com/c/ashrae-energy-prediction/discussion/115698#latest-674237) that **site 4 is UC Berkeley**. Fortunately, the university provides its energry consumption data [publicly](https://engagementdashboard.com/ucb/ucb/). \n\nThe following kernel extracts the historical energy consumption data for several buildings. Even though the dashboard shows one year data maximum, rest service returns data from 2016.\n\nAs you might realize, some buildings have same square feet and built year values in both ASHRAE and Berkeley data set. Besides, some buildings have very close square feet area and same built year. There are **49 buildings** in Berkeley correlated to ASHRAE data set.\n\n[https://www.kaggle.com/serengil/ucb-data-leakage-site-4](https://www.kaggle.com/serengil/ucb-data-leakage-site-4)", "677951": "Potentially, I am starting to suspect that all sites might be publicly collected at this point. :( Is it intentional to screw up the nature of kaggle?\n━━━━━┓\n┓┓┓┓┓┃/ . `(와장창)\n┓┓┓┓┓ : . ＼○ノ\n\n┓┓┓┓┓┃ ` /\n\n┓┓┓┓┓┃ ノ) .\n┓┓┓┓┓┃ ,\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃\n┓┓┓┓┓┃", "680423": "Initial kernel I've published found 49 buildings but currently 74 buildings exactly matched in both UC Berkeley and ASHRAE data sets. This is equal to **1.286.941** instances in submission. Notice that submission set was **41.697.600** instances. This means that site 4 leak data is **3.08%** of all submission data." }, "source": "meta" }