Apologies in advance for basic and late nature of my question. I am new to Kaggle/Python and I just joined this competition. In general I see the \"Overview\" has some intro to competition, but I am having trouble finding info to get started in terms of basic tools specific to the competition. For example, what are some of the basic functions to load and process dicom data? I ask, because I see how basic the \"Overview\" is with respect to other details like background of tumor detection in the brain, data sources and acknowledgements etc. but not a lot on formal bits and pieces of the actual competition problem.
\nI do see that by going through the discussion posts it's possible to come across pieces of the basic data load and processing components, modules like the pydicom for example. But is there a specific section where some of these basic tools might be introduced for the competition, one that I might be over-looking?
\nThank you.
Apologies in advance for basic and late nature of my question. I am new to Kaggle/Python and I just joined this competition. In general I see the \"Overview\" has some intro to competition, but I am having trouble finding info to get started in terms of basic tools specific to the competition. For example, what are some of the basic functions to load and process dicom data? I ask, because I see how basic the \"Overview\" is with respect to other details like background of tumor detection in the brain, data sources and acknowledgements etc. but not a lot on formal bits and pieces of the actual competition problem.
\nI do see that by going through the discussion posts it's possible to come across pieces of the basic data load and processing components, modules like the pydicom for example. But is there a specific section where some of these basic tools might be introduced for the competition, one that I might be over-looking?
\nThank you.
HI @nanocipher . I don't think the Overview section will cover how you can load the data to start preparing the model. The reason I can think of is:
\nIt is a part of the Data Pipeline, everyone uses a different approach to prepare the pipeline. Hence it is very hard for the hosts to give one general process.
\nAlso, every data is of different nature, and there are more than one ways to load it.
However, you can always keep an eye on the discussion forums & code, where outstanding folks of the community come up with nice ideas/hacks on how to start with the competition.
\nFor example, you can take a look at one of these notebooks to have an idea:
\nYou can keep an eye on tags like visualization, EDA, starter-notebook, etc.
\nHope this helps, have a great day!
", "rawMarkdown": "HI @nanocipher . I don't think the Overview section will cover how you can load the data to start preparing the model. The reason I can think of is:\nIt is a part of the Data Pipeline, everyone uses a different approach to prepare the pipeline. Hence it is very hard for the hosts to give one general process.\nAlso, every data is of different nature, and there are more than one ways to load it.\n\nHowever, you can always keep an eye on the discussion forums & code, where outstanding folks of the community come up with nice ideas/hacks on how to start with the competition.\n\nFor example, you can take a look at one of these notebooks to have an idea:\n1. https://www.kaggle.com/ivgona/rnsa-miccai-glioblastoma-eda\n2. https://www.kaggle.com/gpreda/visualize-ct-dicom-data\n3. https://www.kaggle.com/adkarhe/dicom-images\n\nYou can keep an eye on tags like visualization, EDA, starter-notebook, etc.\n\nHope this helps, have a great day!", "votes": null }, { "id": "1504038", "postDate": "09/06/2021 03:01:52", "content": "There is a video by sentdex where he goes over loading and preprocessing a very similar kaggle dataset. Should be a good starting point
\nhttps://www.kaggle.com/sentdex/first-pass-through-data-w-3d-convnet
This is what i used for understanding and loading the data.
", "rawMarkdown": "There is a video by sentdex where he goes over loading and preprocessing a very similar kaggle dataset. Should be a good starting point \nhttps://www.kaggle.com/sentdex/first-pass-through-data-w-3d-convnet\n\nThis is what i used for understanding and loading the data.", "votes": null }, { "id": "1504698", "postDate": "09/06/2021 15:26:53", "content": "Thank you very much, indeed thats exactly what I was looking for. I wish the organizers of the competition could for example add the links you shared here in the overview or a separate \" useful code/info\" section so everyone can see them. Information is clearly out there, it's just hard to find it without sifting through massive amount of other peripheral data. Something like an overview section holds enormous power in terms of delivering critically empowering informational right into the artery and reaching masses. It would be so helpful if it could be augmented/updated, at organizer's discretion, by referencing posts like these. Fantastic posts. Thank you again.
", "rawMarkdown": "Thank you very much, indeed thats exactly what I was looking for. I wish the organizers of the competition could for example add the links you shared here in the overview or a separate \" useful code/info\" section so everyone can see them. Information is clearly out there, it's just hard to find it without sifting through massive amount of other peripheral data. Something like an overview section holds enormous power in terms of delivering critically empowering informational right into the artery and reaching masses. It would be so helpful if it could be augmented/updated, at organizer's discretion, by referencing posts like these. Fantastic posts. Thank you again.", "votes": null } ], "comments": [ { "id": 1504008, "author_name": "sandynigs", "author_url": "", "post_date": "09/06/2021 01:53:12", "content": "HI @nanocipher . I don't think the Overview section will cover how you can load the data to start preparing the model. The reason I can think of is:
\nIt is a part of the Data Pipeline, everyone uses a different approach to prepare the pipeline. Hence it is very hard for the hosts to give one general process.
\nAlso, every data is of different nature, and there are more than one ways to load it.
However, you can always keep an eye on the discussion forums & code, where outstanding folks of the community come up with nice ideas/hacks on how to start with the competition.
\nFor example, you can take a look at one of these notebooks to have an idea:
\nYou can keep an eye on tags like visualization, EDA, starter-notebook, etc.
\nHope this helps, have a great day!
", "votes": null, "replies": [ { "id": 1504698, "author_name": "nanocipher", "author_url": "", "post_date": "09/06/2021 15:26:53", "content": "Thank you very much, indeed thats exactly what I was looking for. I wish the organizers of the competition could for example add the links you shared here in the overview or a separate \" useful code/info\" section so everyone can see them. Information is clearly out there, it's just hard to find it without sifting through massive amount of other peripheral data. Something like an overview section holds enormous power in terms of delivering critically empowering informational right into the artery and reaching masses. It would be so helpful if it could be augmented/updated, at organizer's discretion, by referencing posts like these. Fantastic posts. Thank you again.
", "votes": null, "replies": [] } ] }, { "id": 1504038, "author_name": "aryamansharma47", "author_url": "", "post_date": "09/06/2021 03:01:52", "content": "There is a video by sentdex where he goes over loading and preprocessing a very similar kaggle dataset. Should be a good starting point
\nhttps://www.kaggle.com/sentdex/first-pass-through-data-w-3d-convnet
This is what i used for understanding and loading the data.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1503937": "Apologies in advance for basic and late nature of my question. I am new to Kaggle/Python and I just joined this competition. In general I see the \"Overview\" has some intro to competition, but I am having trouble finding info to get started in terms of basic tools specific to the competition. For example, what are some of the basic functions to load and process dicom data? I ask, because I see how basic the \"Overview\" is with respect to other details like background of tumor detection in the brain, data sources and acknowledgements etc. but not a lot on formal bits and pieces of the actual competition problem.\n\nI do see that by going through the discussion posts it's possible to come across pieces of the basic data load and processing components, modules like the pydicom for example. But is there a specific section where some of these basic tools might be introduced for the competition, one that I might be over-looking? \n\nThank you.", "1504008": "HI @nanocipher . I don't think the Overview section will cover how you can load the data to start preparing the model. The reason I can think of is:\nIt is a part of the Data Pipeline, everyone uses a different approach to prepare the pipeline. Hence it is very hard for the hosts to give one general process.\nAlso, every data is of different nature, and there are more than one ways to load it.\n\nHowever, you can always keep an eye on the discussion forums & code, where outstanding folks of the community come up with nice ideas/hacks on how to start with the competition.\n\nFor example, you can take a look at one of these notebooks to have an idea:\n1. https://www.kaggle.com/ivgona/rnsa-miccai-glioblastoma-eda\n2. https://www.kaggle.com/gpreda/visualize-ct-dicom-data\n3. https://www.kaggle.com/adkarhe/dicom-images\n\nYou can keep an eye on tags like visualization, EDA, starter-notebook, etc.\n\nHope this helps, have a great day!", "1504038": "There is a video by sentdex where he goes over loading and preprocessing a very similar kaggle dataset. Should be a good starting point \nhttps://www.kaggle.com/sentdex/first-pass-through-data-w-3d-convnet\n\nThis is what i used for understanding and loading the data.", "1504698": "Thank you very much, indeed thats exactly what I was looking for. I wish the organizers of the competition could for example add the links you shared here in the overview or a separate \" useful code/info\" section so everyone can see them. Information is clearly out there, it's just hard to find it without sifting through massive amount of other peripheral data. Something like an overview section holds enormous power in terms of delivering critically empowering informational right into the artery and reaching masses. It would be so helpful if it could be augmented/updated, at organizer's discretion, by referencing posts like these. Fantastic posts. Thank you again." }, "source": "meta" }