{ "id": 310867, "title": "Taking into account if a product is still available", "url": "/competitions/h-and-m-personalized-fashion-recommendations/discussion/310867", "author_name": "", "post_date": "2022-03-03T13:44:42.579246200Z", "votes": 13, "comment_count": 7, "views": 0, "content": "

Hi everybody,
\nI was wondering about one characteristic of the fashion market: it’s speed. H&M is selling fast fashion. Items that were available three month ago, might no longer be available today. While this might not be true for basic stuff like socks or black underwear, it certainly is for the „trendy stuff“.

\n

How do you handle this in your recommendations?
\nIs anybody checking if a product might be out of stock already and excluding this product from recommendations?

\n

Any thoughts on this topic are very welcome.

", "messages": [ { "id": "1710930", "postDate": "03/03/2022 13:44:42", "content": "

Hi everybody,
\nI was wondering about one characteristic of the fashion market: it’s speed. H&M is selling fast fashion. Items that were available three month ago, might no longer be available today. While this might not be true for basic stuff like socks or black underwear, it certainly is for the „trendy stuff“.

\n

How do you handle this in your recommendations?
\nIs anybody checking if a product might be out of stock already and excluding this product from recommendations?

\n

Any thoughts on this topic are very welcome.

", "rawMarkdown": "Hi everybody, \nI was wondering about one characteristic of the fashion market: it’s speed. H&M is selling fast fashion. Items that were available three month ago, might no longer be available today. While this might not be true for basic stuff like socks or black underwear, it certainly is for the „trendy stuff“. \n\nHow do you handle this in your recommendations? \nIs anybody checking if a product might be out of stock already and excluding this product from recommendations?\n\nAny thoughts on this topic are very welcome.", "votes": null }, { "id": "1711862", "postDate": "03/04/2022 11:37:11", "content": "

I think we can develop an additional attribute about some of time series characteristics, e.g. last sales date of an item, or if an item is in declining trend. I did a bit of this on my notebook: https://www.kaggle.com/aussie84/product-profiling-and-price-trend-analysis

\n

But since it was for EDA, I only looked at Top 10 articles for each year. There are some items that are very popular in 2018, but it just disappeared away.
\n\"\"

\n

But some things like black socks are just always best selling :D
\n\"\"

\n

Nonetheless, still haven't looked into proper way to scale this across all articles (unless just taking last sales date, or aggregate the 2020 sales vs. threshold)

", "rawMarkdown": "I think we can develop an additional attribute about some of time series characteristics, e.g. last sales date of an item, or if an item is in declining trend. I did a bit of this on my notebook: https://www.kaggle.com/aussie84/product-profiling-and-price-trend-analysis\n\nBut since it was for EDA, I only looked at Top 10 articles for each year. There are some items that are very popular in 2018, but it just disappeared away.\n![](https://www.kaggleusercontent.com/kf/88295392/eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..zfMnDJuFTn4gRv38r0QyFw.Bukr6iGWV2bz-2lIT-d3gRnG5L61ENjIWv1eSEoHlv0TbqtGYSiwe4bCIIe_FfVCEJh02rc5aKWtT5pr8QfJtSmj2IvvBsX87AcVLpE-e6X4dtxeatmooRiAxibpHWxyUmFm1QkI4aRR2WhQZarc8ZdNZPso9OTqXW1C9pdyVp2bA3kJXh24G1UcrYr6T-6LDeaYlLkXOsgrBqAhpaLkAtO8usxYkI7vRgfW4oz8ndG_69hygfKc2sXjF1MRjhfrw1MTxypem9wpJzBCuIz3zR_3k0IIAWkIi65CBRCAK_-WKWyeB3cb8vEbS1oZB_uybnDKXUNra3E64Oxl28sfUSP0mnojv4hCDJn5BLJTITtYPz8xxyM2SThstwoeY_7fYmLQK7toXsek6b9GPq_qUvo9sG9SZJ_fiQW66YfUDC3wq4pYYSWiTAxnlqIt4gLLlqCLBtCiriyrasjHRUQ5JbUwqtnonaBwX4gIXFOQLeKhYR0dbudXOqaDr2sl-Nmjtv2IEtzqB2WTFVp2vbtr_AyTTD95xzgu8HE_JpM1BuJdEhdJX4PQdmuoUlkeJYiAxw16oRslFljkwGXjjOHVNYNboC8YzFQP9cAP7lD9tNaENQXpHL9oGptMZah5BXlVbIgUIx3T8FLncfptBs-vCErQDnae7AJoFXfxqkfcPQGinaKfNx7zfMFeX_LCd19r.OGoHN5pCLZQ1na_jCoCNiQ/__results___files/__results___31_1.png)\n\nBut some things like black socks are just always best selling :D\n![](https://www.kaggleusercontent.com/kf/88295392/eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..zfMnDJuFTn4gRv38r0QyFw.Bukr6iGWV2bz-2lIT-d3gRnG5L61ENjIWv1eSEoHlv0TbqtGYSiwe4bCIIe_FfVCEJh02rc5aKWtT5pr8QfJtSmj2IvvBsX87AcVLpE-e6X4dtxeatmooRiAxibpHWxyUmFm1QkI4aRR2WhQZarc8ZdNZPso9OTqXW1C9pdyVp2bA3kJXh24G1UcrYr6T-6LDeaYlLkXOsgrBqAhpaLkAtO8usxYkI7vRgfW4oz8ndG_69hygfKc2sXjF1MRjhfrw1MTxypem9wpJzBCuIz3zR_3k0IIAWkIi65CBRCAK_-WKWyeB3cb8vEbS1oZB_uybnDKXUNra3E64Oxl28sfUSP0mnojv4hCDJn5BLJTITtYPz8xxyM2SThstwoeY_7fYmLQK7toXsek6b9GPq_qUvo9sG9SZJ_fiQW66YfUDC3wq4pYYSWiTAxnlqIt4gLLlqCLBtCiriyrasjHRUQ5JbUwqtnonaBwX4gIXFOQLeKhYR0dbudXOqaDr2sl-Nmjtv2IEtzqB2WTFVp2vbtr_AyTTD95xzgu8HE_JpM1BuJdEhdJX4PQdmuoUlkeJYiAxw16oRslFljkwGXjjOHVNYNboC8YzFQP9cAP7lD9tNaENQXpHL9oGptMZah5BXlVbIgUIx3T8FLncfptBs-vCErQDnae7AJoFXfxqkfcPQGinaKfNx7zfMFeX_LCd19r.OGoHN5pCLZQ1na_jCoCNiQ/__results___files/__results___31_9.png)\n\nNonetheless, still haven't looked into proper way to scale this across all articles (unless just taking last sales date, or aggregate the 2020 sales vs. threshold)", "votes": null }, { "id": "1712565", "postDate": "03/05/2022 03:59:33", "content": "

@melanie7744

\n

Most of the public notebooks seems to be based on recently sold and frequently sold together items. So if a product was sold recently, it would have been available even if not new as you pointed out for socks or underwear.

\n

Does that make sense?

\n

You can look at last sale date or items not sold at all ever to filter your article catalog also.

", "rawMarkdown": "melanie7744 \n\nMost of the public notebooks seems to be based on recently sold and frequently sold together items. So if a product was sold recently, it would have been available even if not new as you pointed out for socks or underwear.\n\nDoes that make sense?\n\nYou can look at last sale date or items not sold at all ever to filter your article catalog also.", "votes": null }, { "id": "1712851", "postDate": "03/05/2022 12:18:53", "content": "

Thanks for your input! Filtering the article catalog by last sales date seems like a good idea to prevent recommending outdated items.

", "rawMarkdown": "Thanks for your input! Filtering the article catalog by last sales date seems like a good idea to prevent recommending outdated items.", "votes": null }, { "id": "1712853", "postDate": "03/05/2022 12:19:52", "content": "

I'll take a look at your notebook. Especially at the popular items that disappear. Thanks for sharing!

", "rawMarkdown": "I'll take a look at your notebook. Especially at the popular items that disappear. Thanks for sharing!", "votes": null }, { "id": "1712938", "postDate": "03/05/2022 13:50:05", "content": "

@melanie7744 , @aussie84
\nMaybe this notebook from @byfone will help https://www.kaggle.com/byfone/h-m-trending-products-weekly

", "rawMarkdown": "melanie7744 , @aussie84 \nMaybe this notebook from @byfone will help https://www.kaggle.com/byfone/h-m-trending-products-weekly", "votes": null }, { "id": "1713554", "postDate": "03/06/2022 06:28:18", "content": "

nice reference, thanks!

", "rawMarkdown": "nice reference, thanks!", "votes": null }, { "id": "1718763", "postDate": "03/11/2022 06:20:02", "content": "

@melanie7744
\nMaybe this notebook help to you!!
\nhttps://www.kaggle.com/pyy0715/let-s-predict-the-articles-with-the-will-be-sold

", "rawMarkdown": "melanie7744 \nMaybe this notebook help to you!!\nhttps://www.kaggle.com/pyy0715/let-s-predict-the-articles-with-the-will-be-sold", "votes": null } ], "comments": [ { "id": 1711862, "author_name": "aussie84", "author_url": "", "post_date": "03/04/2022 11:37:11", "content": "

I think we can develop an additional attribute about some of time series characteristics, e.g. last sales date of an item, or if an item is in declining trend. I did a bit of this on my notebook: https://www.kaggle.com/aussie84/product-profiling-and-price-trend-analysis

\n

But since it was for EDA, I only looked at Top 10 articles for each year. There are some items that are very popular in 2018, but it just disappeared away.
\n\"\"

\n

But some things like black socks are just always best selling :D
\n\"\"

\n

Nonetheless, still haven't looked into proper way to scale this across all articles (unless just taking last sales date, or aggregate the 2020 sales vs. threshold)

", "votes": null, "replies": [ { "id": 1712853, "author_name": "melanie7744", "author_url": "", "post_date": "03/05/2022 12:19:52", "content": "

I'll take a look at your notebook. Especially at the popular items that disappear. Thanks for sharing!

", "votes": null, "replies": [] } ] }, { "id": 1712565, "author_name": "atulverma", "author_url": "", "post_date": "03/05/2022 03:59:33", "content": "

@melanie7744

\n

Most of the public notebooks seems to be based on recently sold and frequently sold together items. So if a product was sold recently, it would have been available even if not new as you pointed out for socks or underwear.

\n

Does that make sense?

\n

You can look at last sale date or items not sold at all ever to filter your article catalog also.

", "votes": null, "replies": [ { "id": 1712851, "author_name": "melanie7744", "author_url": "", "post_date": "03/05/2022 12:18:53", "content": "

Thanks for your input! Filtering the article catalog by last sales date seems like a good idea to prevent recommending outdated items.

", "votes": null, "replies": [] } ] }, { "id": 1712938, "author_name": "atulverma", "author_url": "", "post_date": "03/05/2022 13:50:05", "content": "

@melanie7744 , @aussie84
\nMaybe this notebook from @byfone will help https://www.kaggle.com/byfone/h-m-trending-products-weekly

", "votes": null, "replies": [ { "id": 1713554, "author_name": "aussie84", "author_url": "", "post_date": "03/06/2022 06:28:18", "content": "

nice reference, thanks!

", "votes": null, "replies": [] } ] }, { "id": 1718763, "author_name": "pyy0715", "author_url": "", "post_date": "03/11/2022 06:20:02", "content": "

@melanie7744
\nMaybe this notebook help to you!!
\nhttps://www.kaggle.com/pyy0715/let-s-predict-the-articles-with-the-will-be-sold

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1710930": "Hi everybody, \nI was wondering about one characteristic of the fashion market: it’s speed. H&M is selling fast fashion. Items that were available three month ago, might no longer be available today. While this might not be true for basic stuff like socks or black underwear, it certainly is for the „trendy stuff“. \n\nHow do you handle this in your recommendations? \nIs anybody checking if a product might be out of stock already and excluding this product from recommendations?\n\nAny thoughts on this topic are very welcome.", "1711862": "I think we can develop an additional attribute about some of time series characteristics, e.g. last sales date of an item, or if an item is in declining trend. I did a bit of this on my notebook: https://www.kaggle.com/aussie84/product-profiling-and-price-trend-analysis\n\nBut since it was for EDA, I only looked at Top 10 articles for each year. There are some items that are very popular in 2018, but it just disappeared away.\n![](https://www.kaggleusercontent.com/kf/88295392/eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..zfMnDJuFTn4gRv38r0QyFw.Bukr6iGWV2bz-2lIT-d3gRnG5L61ENjIWv1eSEoHlv0TbqtGYSiwe4bCIIe_FfVCEJh02rc5aKWtT5pr8QfJtSmj2IvvBsX87AcVLpE-e6X4dtxeatmooRiAxibpHWxyUmFm1QkI4aRR2WhQZarc8ZdNZPso9OTqXW1C9pdyVp2bA3kJXh24G1UcrYr6T-6LDeaYlLkXOsgrBqAhpaLkAtO8usxYkI7vRgfW4oz8ndG_69hygfKc2sXjF1MRjhfrw1MTxypem9wpJzBCuIz3zR_3k0IIAWkIi65CBRCAK_-WKWyeB3cb8vEbS1oZB_uybnDKXUNra3E64Oxl28sfUSP0mnojv4hCDJn5BLJTITtYPz8xxyM2SThstwoeY_7fYmLQK7toXsek6b9GPq_qUvo9sG9SZJ_fiQW66YfUDC3wq4pYYSWiTAxnlqIt4gLLlqCLBtCiriyrasjHRUQ5JbUwqtnonaBwX4gIXFOQLeKhYR0dbudXOqaDr2sl-Nmjtv2IEtzqB2WTFVp2vbtr_AyTTD95xzgu8HE_JpM1BuJdEhdJX4PQdmuoUlkeJYiAxw16oRslFljkwGXjjOHVNYNboC8YzFQP9cAP7lD9tNaENQXpHL9oGptMZah5BXlVbIgUIx3T8FLncfptBs-vCErQDnae7AJoFXfxqkfcPQGinaKfNx7zfMFeX_LCd19r.OGoHN5pCLZQ1na_jCoCNiQ/__results___files/__results___31_1.png)\n\nBut some things like black socks are just always best selling :D\n![](https://www.kaggleusercontent.com/kf/88295392/eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2In0..zfMnDJuFTn4gRv38r0QyFw.Bukr6iGWV2bz-2lIT-d3gRnG5L61ENjIWv1eSEoHlv0TbqtGYSiwe4bCIIe_FfVCEJh02rc5aKWtT5pr8QfJtSmj2IvvBsX87AcVLpE-e6X4dtxeatmooRiAxibpHWxyUmFm1QkI4aRR2WhQZarc8ZdNZPso9OTqXW1C9pdyVp2bA3kJXh24G1UcrYr6T-6LDeaYlLkXOsgrBqAhpaLkAtO8usxYkI7vRgfW4oz8ndG_69hygfKc2sXjF1MRjhfrw1MTxypem9wpJzBCuIz3zR_3k0IIAWkIi65CBRCAK_-WKWyeB3cb8vEbS1oZB_uybnDKXUNra3E64Oxl28sfUSP0mnojv4hCDJn5BLJTITtYPz8xxyM2SThstwoeY_7fYmLQK7toXsek6b9GPq_qUvo9sG9SZJ_fiQW66YfUDC3wq4pYYSWiTAxnlqIt4gLLlqCLBtCiriyrasjHRUQ5JbUwqtnonaBwX4gIXFOQLeKhYR0dbudXOqaDr2sl-Nmjtv2IEtzqB2WTFVp2vbtr_AyTTD95xzgu8HE_JpM1BuJdEhdJX4PQdmuoUlkeJYiAxw16oRslFljkwGXjjOHVNYNboC8YzFQP9cAP7lD9tNaENQXpHL9oGptMZah5BXlVbIgUIx3T8FLncfptBs-vCErQDnae7AJoFXfxqkfcPQGinaKfNx7zfMFeX_LCd19r.OGoHN5pCLZQ1na_jCoCNiQ/__results___files/__results___31_9.png)\n\nNonetheless, still haven't looked into proper way to scale this across all articles (unless just taking last sales date, or aggregate the 2020 sales vs. threshold)", "1712565": "melanie7744 \n\nMost of the public notebooks seems to be based on recently sold and frequently sold together items. So if a product was sold recently, it would have been available even if not new as you pointed out for socks or underwear.\n\nDoes that make sense?\n\nYou can look at last sale date or items not sold at all ever to filter your article catalog also.", "1712851": "Thanks for your input! Filtering the article catalog by last sales date seems like a good idea to prevent recommending outdated items.", "1712853": "I'll take a look at your notebook. Especially at the popular items that disappear. Thanks for sharing!", "1712938": "melanie7744 , @aussie84 \nMaybe this notebook from @byfone will help https://www.kaggle.com/byfone/h-m-trending-products-weekly", "1713554": "nice reference, thanks!", "1718763": "melanie7744 \nMaybe this notebook help to you!!\nhttps://www.kaggle.com/pyy0715/let-s-predict-the-articles-with-the-will-be-sold" }, "source": "meta" }