{ "id": 393852, "title": "Understanding The Data", "url": "/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/393852", "author_name": "", "post_date": "2023-03-11T03:05:12.842556900Z", "votes": 61, "comment_count": 11, "views": 0, "content": "

Hopefully this post will help give you some clarity about the data.

\n

The dataset is comprised of lower-back 3D accelerometer data. We are provided time series data of each subject. At each time step, we are provided 3 measurements:

\n\n

Here is a helpful visualization:
\n\"\"

\n

Example of one files measurements plotted:
\n\"\"

", "messages": [ { "id": "2176905", "postDate": "03/11/2023 03:05:12", "content": "

Hopefully this post will help give you some clarity about the data.

\n

The dataset is comprised of lower-back 3D accelerometer data. We are provided time series data of each subject. At each time step, we are provided 3 measurements:

\n\n

Here is a helpful visualization:
\n\"\"

\n

Example of one files measurements plotted:
\n\"\"

", "rawMarkdown": "Hopefully this post will help give you some clarity about the data.\n\nThe dataset is comprised of lower-back 3D accelerometer data. We are provided time series data of each subject. At each time step, we are provided 3 measurements: \n- **AccV** - acceleration from vertical sensor\n- **AccML** - acceleration from mediolateral sensor\n- **AccAP** - acceleration from anteroposterior sensor\n\nHere is a helpful visualization:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6537187%2F7ee843f278c9ee3b0b598905e38dd903%2FDirections-of-displacement-AP-Anterior-posterior-direction-forward-backward.png?generation=1678503484530383&alt=media)\n\nExample of one files measurements plotted:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6537187%2F701336f11a01e438007d7e6de2989b20%2FV_ML_AP_Ex.png?generation=1678503888698848&alt=media)", "votes": null }, { "id": "2177833", "postDate": "03/11/2023 20:33:50", "content": "

Very good approach. Is that chart above plotted by you?

\n

If not, don't forget to provide credits. Never forget author's rights Shah.

", "rawMarkdown": "Very good approach. Is that chart above plotted by you?\n\nIf not, don't forget to provide credits. Never forget author's rights Shah.", "votes": null }, { "id": "2177886", "postDate": "03/11/2023 22:08:15", "content": "

@mpwolke
\nYes I created the chart. Here is the code:

\n
df = pd.read_csv()\nplt.plot(df.AccV, color=, label=)\nplt.plot(df.AccML, color=, label=)\nplt.plot(df.AccAP, color=, label=)\nplt.xlabel()\nplt.ylabel()\nplt.legend()\nplt.show()\n
\n

The other figure from here https://www.researchgate.net/figure/Directions-of-displacement-AP-Anterior-posterior-direction-forward-backward_fig6_225059668

", "rawMarkdown": "mpwolke \nYes I created the chart. Here is the code:\n```python\ndf = pd.read_csv(\"/kaggle/input/tlvmc-parkinsons-freezing-gait-prediction/train/defog/02ea782681.csv\")\nplt.plot(df.AccV, color='#69cf83', label='V')\nplt.plot(df.AccML, color='#d6b258', label='ML')\nplt.plot(df.AccAP, color='#96bcfa', label='AP')\nplt.xlabel('Time')\nplt.ylabel('Sensor Value')\nplt.legend()\nplt.show()\n```\n\nThe other figure from here https://www.researchgate.net/figure/Directions-of-displacement-AP-Anterior-posterior-direction-forward-backward_fig6_225059668", "votes": null }, { "id": "2177910", "postDate": "03/11/2023 22:24:00", "content": "

Thank you for clarifying it. I asked because I couldn't find any public Notebook by you till now. I'll be glad to read it.

\n

Valuable approach since the accelerometers would have great importance for the prediction (that was endorsed by Kitamura).

", "rawMarkdown": "Thank you for clarifying it. I asked because I couldn't find any public Notebook by you till now. I'll be glad to read it.\n\nValuable approach since the accelerometers would have great importance for the prediction (that was endorsed by Kitamura).", "votes": null }, { "id": "2180287", "postDate": "03/13/2023 17:44:24", "content": "

I made the notebook public with a bunch of other EDA info here

", "rawMarkdown": "I made the notebook public with a bunch of other EDA info [here](https://www.kaggle.com/code/ravishah1/parkinson-s-fog-accelerometer-data-eda)", "votes": null }, { "id": "2180343", "postDate": "03/13/2023 18:21:44", "content": "

Been there. You did it great Ravi.

", "rawMarkdown": "Been there. You did it great Ravi.", "votes": null }, { "id": "2181136", "postDate": "03/14/2023 10:57:20", "content": "

This is a useful explanation of the dataset for anyone working with the lower-back 3D accelerometer data. It provides a clear understanding of what the data contains, including the measurements for AccV, AccML, and AccAP. Having a good grasp of the data is crucial for making informed decisions in data analysis and modeling.

", "rawMarkdown": "This is a useful explanation of the dataset for anyone working with the lower-back 3D accelerometer data. It provides a clear understanding of what the data contains, including the measurements for AccV, AccML, and AccAP. Having a good grasp of the data is crucial for making informed decisions in data analysis and modeling.", "votes": null }, { "id": "2183325", "postDate": "03/15/2023 15:33:53", "content": "

I'm glad you found it useful @tariqbashir!

", "rawMarkdown": "I'm glad you found it useful @tariqbashir!", "votes": null }, { "id": "2190728", "postDate": "03/21/2023 12:58:29", "content": "

@ravishah1 thanks for sharing. What’s the time resolution of the acc data?

", "rawMarkdown": "ravishah1 thanks for sharing. What’s the time resolution of the acc data?", "votes": null }, { "id": "2191429", "postDate": "03/22/2023 01:07:23", "content": "

To quote the documentation in the competition's data tab \"Series from the tdcsfog dataset are recorded at 128Hz (128 timesteps per second), while series from the defog and daily series are recorded at 100Hz (100 timesteps per second).\"

", "rawMarkdown": "To quote the documentation in the competition's data tab \"Series from the tdcsfog dataset are recorded at 128Hz (128 timesteps per second), while series from the defog and daily series are recorded at 100Hz (100 timesteps per second).\"", "votes": null }, { "id": "2243579", "postDate": "05/03/2023 02:30:11", "content": "

Thanks for the excellent, concise and lucid explanation - here is the code for the function to display the 3D sensor line chart.

\n
\n ():\n    plt.figure()\n    plt.plot(df[], df[], label=, linewidth=)\n    plt.plot(df[], df[], label=, linewidth=)\n    plt.plot(df[], df[], label=, linewidth=)\n\n    plt.xlabel()\n    plt.ylabel()\n    plt.title()\n    plt.legend()\n    plt.show()\n
", "rawMarkdown": "Thanks for the excellent, concise and lucid explanation - here is the code for the function to display the 3D sensor line chart.\n```python\n# Define a function to display a line chart of 3D accelerometer data \ndef plot_acc_graph(df):\n plt.figure()\n plt.plot(df[\"Time\"], df[\"AccV\"], label=\"AccV\", linewidth=0.7)\n plt.plot(df[\"Time\"], df[\"AccML\"], label=\"AccML\", linewidth=0.7)\n plt.plot(df[\"Time\"], df[\"AccAP\"], label=\"AccAP\", linewidth=0.7)\n\n plt.xlabel(\"Time\")\n plt.ylabel(\"Acceleration\")\n plt.title(\"Acceleration vs Time\")\n plt.legend()\n plt.show()\n```", "votes": null }, { "id": "2294030", "postDate": "06/09/2023 17:05:19", "content": "

Great content here! Thanks @ravishah1 🙏

", "rawMarkdown": "Great content here! Thanks @ravishah1 🙏", "votes": null } ], "comments": [ { "id": 2177833, "author_name": "mpwolke", "author_url": "", "post_date": "03/11/2023 20:33:50", "content": "

Very good approach. Is that chart above plotted by you?

\n

If not, don't forget to provide credits. Never forget author's rights Shah.

", "votes": null, "replies": [ { "id": 2177886, "author_name": "ravishah1", "author_url": "", "post_date": "03/11/2023 22:08:15", "content": "

@mpwolke
\nYes I created the chart. Here is the code:

\n
df = pd.read_csv()\nplt.plot(df.AccV, color=, label=)\nplt.plot(df.AccML, color=, label=)\nplt.plot(df.AccAP, color=, label=)\nplt.xlabel()\nplt.ylabel()\nplt.legend()\nplt.show()\n
\n

The other figure from here https://www.researchgate.net/figure/Directions-of-displacement-AP-Anterior-posterior-direction-forward-backward_fig6_225059668

", "votes": null, "replies": [ { "id": 2177910, "author_name": "mpwolke", "author_url": "", "post_date": "03/11/2023 22:24:00", "content": "

Thank you for clarifying it. I asked because I couldn't find any public Notebook by you till now. I'll be glad to read it.

\n

Valuable approach since the accelerometers would have great importance for the prediction (that was endorsed by Kitamura).

", "votes": null, "replies": [] } ] }, { "id": 2180287, "author_name": "ravishah1", "author_url": "", "post_date": "03/13/2023 17:44:24", "content": "

I made the notebook public with a bunch of other EDA info here

", "votes": null, "replies": [ { "id": 2180343, "author_name": "mpwolke", "author_url": "", "post_date": "03/13/2023 18:21:44", "content": "

Been there. You did it great Ravi.

", "votes": null, "replies": [] } ] } ] }, { "id": 2181136, "author_name": "tariqbashir", "author_url": "", "post_date": "03/14/2023 10:57:20", "content": "

This is a useful explanation of the dataset for anyone working with the lower-back 3D accelerometer data. It provides a clear understanding of what the data contains, including the measurements for AccV, AccML, and AccAP. Having a good grasp of the data is crucial for making informed decisions in data analysis and modeling.

", "votes": null, "replies": [ { "id": 2183325, "author_name": "ravishah1", "author_url": "", "post_date": "03/15/2023 15:33:53", "content": "

I'm glad you found it useful @tariqbashir!

", "votes": null, "replies": [] } ] }, { "id": 2190728, "author_name": "joshuascj", "author_url": "", "post_date": "03/21/2023 12:58:29", "content": "

@ravishah1 thanks for sharing. What’s the time resolution of the acc data?

", "votes": null, "replies": [ { "id": 2191429, "author_name": "jsday96", "author_url": "", "post_date": "03/22/2023 01:07:23", "content": "

To quote the documentation in the competition's data tab \"Series from the tdcsfog dataset are recorded at 128Hz (128 timesteps per second), while series from the defog and daily series are recorded at 100Hz (100 timesteps per second).\"

", "votes": null, "replies": [] } ] }, { "id": 2243579, "author_name": "tomonorisasaki", "author_url": "", "post_date": "05/03/2023 02:30:11", "content": "

Thanks for the excellent, concise and lucid explanation - here is the code for the function to display the 3D sensor line chart.

\n
\n ():\n    plt.figure()\n    plt.plot(df[], df[], label=, linewidth=)\n    plt.plot(df[], df[], label=, linewidth=)\n    plt.plot(df[], df[], label=, linewidth=)\n\n    plt.xlabel()\n    plt.ylabel()\n    plt.title()\n    plt.legend()\n    plt.show()\n
", "votes": null, "replies": [] }, { "id": 2294030, "author_name": "vladiluzjr", "author_url": "", "post_date": "06/09/2023 17:05:19", "content": "

Great content here! Thanks @ravishah1 🙏

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "2176905": "Hopefully this post will help give you some clarity about the data.\n\nThe dataset is comprised of lower-back 3D accelerometer data. We are provided time series data of each subject. At each time step, we are provided 3 measurements: \n- **AccV** - acceleration from vertical sensor\n- **AccML** - acceleration from mediolateral sensor\n- **AccAP** - acceleration from anteroposterior sensor\n\nHere is a helpful visualization:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6537187%2F7ee843f278c9ee3b0b598905e38dd903%2FDirections-of-displacement-AP-Anterior-posterior-direction-forward-backward.png?generation=1678503484530383&alt=media)\n\nExample of one files measurements plotted:\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F6537187%2F701336f11a01e438007d7e6de2989b20%2FV_ML_AP_Ex.png?generation=1678503888698848&alt=media)", "2177833": "Very good approach. Is that chart above plotted by you?\n\nIf not, don't forget to provide credits. Never forget author's rights Shah.", "2177886": "mpwolke \nYes I created the chart. Here is the code:\n```python\ndf = pd.read_csv(\"/kaggle/input/tlvmc-parkinsons-freezing-gait-prediction/train/defog/02ea782681.csv\")\nplt.plot(df.AccV, color='#69cf83', label='V')\nplt.plot(df.AccML, color='#d6b258', label='ML')\nplt.plot(df.AccAP, color='#96bcfa', label='AP')\nplt.xlabel('Time')\nplt.ylabel('Sensor Value')\nplt.legend()\nplt.show()\n```\n\nThe other figure from here https://www.researchgate.net/figure/Directions-of-displacement-AP-Anterior-posterior-direction-forward-backward_fig6_225059668", "2177910": "Thank you for clarifying it. I asked because I couldn't find any public Notebook by you till now. I'll be glad to read it.\n\nValuable approach since the accelerometers would have great importance for the prediction (that was endorsed by Kitamura).", "2180287": "I made the notebook public with a bunch of other EDA info [here](https://www.kaggle.com/code/ravishah1/parkinson-s-fog-accelerometer-data-eda)", "2180343": "Been there. You did it great Ravi.", "2181136": "This is a useful explanation of the dataset for anyone working with the lower-back 3D accelerometer data. It provides a clear understanding of what the data contains, including the measurements for AccV, AccML, and AccAP. Having a good grasp of the data is crucial for making informed decisions in data analysis and modeling.", "2183325": "I'm glad you found it useful @tariqbashir!", "2190728": "ravishah1 thanks for sharing. What’s the time resolution of the acc data?", "2191429": "To quote the documentation in the competition's data tab \"Series from the tdcsfog dataset are recorded at 128Hz (128 timesteps per second), while series from the defog and daily series are recorded at 100Hz (100 timesteps per second).\"", "2243579": "Thanks for the excellent, concise and lucid explanation - here is the code for the function to display the 3D sensor line chart.\n```python\n# Define a function to display a line chart of 3D accelerometer data \ndef plot_acc_graph(df):\n plt.figure()\n plt.plot(df[\"Time\"], df[\"AccV\"], label=\"AccV\", linewidth=0.7)\n plt.plot(df[\"Time\"], df[\"AccML\"], label=\"AccML\", linewidth=0.7)\n plt.plot(df[\"Time\"], df[\"AccAP\"], label=\"AccAP\", linewidth=0.7)\n\n plt.xlabel(\"Time\")\n plt.ylabel(\"Acceleration\")\n plt.title(\"Acceleration vs Time\")\n plt.legend()\n plt.show()\n```", "2294030": "Great content here! Thanks @ravishah1 🙏" }, "source": "meta" }