![\"cc68f19b708c\"](\"https://www.kaggleusercontent.com/kf/131430328/eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..A8-jPWxUNcmTyabNSbr-OQ.ezFfhUHyKqWH_YWSuAjrgS0twtD1z-iyZqLSPd79qw-mVwN2ku9cVDCBiqkHvWkYTJMQOPOj-XD6qQexLecmolzUCE99YFE7OjTVIegLPISVdqtLM2XNYCvFlO3pvi5Fz_ibddbMWXBBtA88pRpvJdBtztA2eTUyd21djXvb6ib7QR2PaS80XDhhBzsmE1NmfGiYc3nr1TZOly6AJWEvXuMJ44E6CSSuRr2-dEqtbyfY34Tev62WdFthkFQazgcyCAVAvF6KLN-N1FrYYeujiMruG_HMRZrD5mBoI_lkjs_69exZmIaE6MtRg_HnSezavFNna6GSk8xhgkws-S7j3YExmrCbyuR2RPWh4XHu9uPZIqdslDbm2YYyBL4TYOO80HTU0E_yhXNeazPadpTaYwhAPgYaGjpFFoQEcXYQVoW69WdswYgFG7FqBIIqYrv9kf3vyShMzVKCAsrkR50kRkUo-5VUIlmjmoGohZ_DgyVEZ8VWST42A1QbRtaFeAX8AOYS7_Vhx3q1wQRNSsVs_7OVvD_LbdtMsiknCWv9nMBGCccgEqC8NX4NaEv6E0Je4Sd8udHHX30sUu7w_ymQqIe4RWVId6ltbbsYBSjJMh6uV58BIGvJFnYW6SICP7mNGlK_QjrFRL3ZhDw9fQBRpbvcL1XZYoO-I2-MgQ_Lolk.SG1JA0SzNnx8H0gcthCctg/__results___files/__results___5_1.png\")
There are some data with a wrong annotation among extracted data. For example, the y-ticks of the chart are from 20 to -25, but they are all annotated as positive values. The data series are shown the same way so that they indicate wrong values.
\nFilename: cc68f19b708c\n\nx_ticks: ['10%', '20%', '30%', '40%', '50%', '60%', '70%', '80%']\ny_ticks: ['20', '15', '10', '5', '0', '5', '10', '15', '20', '25']\nsource: extracted\nchart-type: line\n
Besides negative values, you can find more examples of other cases in my notebook.
\nGenerated data could have the same issue, of course, but I focused on extracted data because the test data are extracted entirely.
I want to know:
\nJust a few days left, but I look forward to your response!
\n(* Please note that it is a different question from this discussion, which pointed out the bounding box issue.)
", "messages": [ { "id": 2279379, "postDate": "2023-05-29T10:18:57.813Z", "content": "There are some data with a wrong annotation among extracted data. For example, the y-ticks of the chart are from 20 to -25, but they are all annotated as positive values. The data series are shown the same way so that they indicate wrong values.
\nFilename: cc68f19b708c\n\nx_ticks: ['10%', '20%', '30%', '40%', '50%', '60%', '70%', '80%']\ny_ticks: ['20', '15', '10', '5', '0', '5', '10', '15', '20', '25']\nsource: extracted\nchart-type: line\n
Besides negative values, you can find more examples of other cases in my notebook.
\nGenerated data could have the same issue, of course, but I focused on extracted data because the test data are extracted entirely.
I want to know:
\nJust a few days left, but I look forward to your response!
\n(* Please note that it is a different question from this discussion, which pointed out the bounding box issue.)
", "rawMarkdown": "There are some data with a wrong annotation among extracted data. For example, the y-ticks of the chart are from 20 to -25, but they are all annotated as positive values. The data series are shown the same way so that they indicate wrong values.\n\n```\nFilename: cc68f19b708c\n\nx_ticks: ['10%', '20%', '30%', '40%', '50%', '60%', '70%', '80%']\ny_ticks: ['20', '15', '10', '5', '0', '5', '10', '15', '20', '25']\nsource: extracted\nchart-type: line\n```\n\n\n\nBesides negative values, you can find more examples of other cases in [my notebook](https://www.kaggle.com/code/sohyunsim/benetech-data-validation/notebook).\nGenerated data could have the same issue, of course, but I focused on extracted data because the test data are extracted entirely.\n\n\nI want to know:\n- if any data annotated improperly is in the test set,\n- and, in this case, how the data will be evaluated.\n\nJust a few days left, but I look forward to your response!\n\n(* Please note that it is a different question from [this discussion](https://www.kaggle.com/datasets/alejopaullier/benetech-wrong-annotations), which pointed out the bounding box issue.)", "votes": 3 } ], "comments": [], "raw_markdown_by_id": { "2279379": "There are some data with a wrong annotation among extracted data. For example, the y-ticks of the chart are from 20 to -25, but they are all annotated as positive values. The data series are shown the same way so that they indicate wrong values.\n\n```\nFilename: cc68f19b708c\n\nx_ticks: ['10%', '20%', '30%', '40%', '50%', '60%', '70%', '80%']\ny_ticks: ['20', '15', '10', '5', '0', '5', '10', '15', '20', '25']\nsource: extracted\nchart-type: line\n```\n\n\n\nBesides negative values, you can find more examples of other cases in [my notebook](https://www.kaggle.com/code/sohyunsim/benetech-data-validation/notebook).\nGenerated data could have the same issue, of course, but I focused on extracted data because the test data are extracted entirely.\n\n\nI want to know:\n- if any data annotated improperly is in the test set,\n- and, in this case, how the data will be evaluated.\n\nJust a few days left, but I look forward to your response!\n\n(* Please note that it is a different question from [this discussion](https://www.kaggle.com/datasets/alejopaullier/benetech-wrong-annotations), which pointed out the bounding box issue.)" } }