{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Summary of this competition\n\n\n\n#### We predict the abnormalities in COVIS-19 by chest radiographs.\n\n#### In particular, radiographs are classified as negative for pneumonia or typical, uncertain, or atypical of COVID-19.\n\n\n\n\n\n\n","metadata":{}},{"cell_type":"markdown","source":"# About this notebook\n\n* ## EDA especially for duplicates in train data\n \n* ## Making clean data","metadata":{}},{"cell_type":"markdown","source":"# Summary image : \n### In short words, I analyzed the duplicates in train data. By using that results, clearly I merge with train_study_level.csv and train_image_level.csv.\n### You can use the merged result, if you like.\n\n\n-------------Attention-----------\n##### If the description image is hard to see, push the copy and edit. And it makes easier to see(It doesn't take much time to run all.)\n##### Futhermore, you can see it more by expanding in the browser. Alternatively, you can download it from input. \n##### I'm sorry it's difficult to attach clear images, due to the capacity limit of the notebook.\n\n","metadata":{}},{"cell_type":"markdown","source":"\n\n![](https://storage.googleapis.com/kagglesdsdata/datasets/1348626/2244049/Clipboard03.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=gcp-kaggle-com%40kaggle-161607.iam.gserviceaccount.com%2F20210518%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20210518T125153Z&X-Goog-Expires=259199&X-Goog-SignedHeaders=host&X-Goog-Signature=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)\n","metadata":{}},{"cell_type":"markdown","source":"# I'm looking forward to helping you even a little. please upvote/follow, thank you!\n# Also, thank you for those who always upvote.","metadata":{}},{"cell_type":"code","source":"import numpy as np \nimport pandas as pd \nimport os\nfrom tqdm import tqdm\nimport matplotlib.pyplot as plt\nimport pydicom","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 0. Confirmation of files","metadata":{}},{"cell_type":"code","source":"train_st = pd.read_csv(\"../input/siim-covid19-detection/train_study_level.csv\")\ntrain_st.head(3)\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"* id - unique study identifier\n\n----classification for predict----\n\n* Negative for Pneumonia \n* Typical Appearance \n* Indeterminate Appearance  \n* Atypical Appearance  ","metadata":{}},{"cell_type":"code","source":"train_im = pd.read_csv(\"../input/siim-covid19-detection/train_image_level.csv\")\ntrain_im.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"* id - including image file name\n* boxes - bounding boxes in easily-readable dictionary format\n* label - the correct prediction label for the provided bounding boxes","metadata":{}},{"cell_type":"code","source":"sample = pd.read_csv(\"../input/siim-covid19-detection/sample_submission.csv\")\nsample.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 1. Get image path and show image","metadata":{}},{"cell_type":"code","source":"train_path = \"../input/siim-covid19-detection/train\"\ntest_path = \"../input/siim-covid19-detection/test\"","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"trainimlist = []\n\nfor dirname, _, filenames in os.walk(train_path):\n    for filename in filenames:\n        trainimlist.append(os.path.join(dirname, filename))","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"testimlist = []\n\nfor dirname, _, filenames in os.walk(test_path):\n    for filename in filenames:\n        testimlist.append(os.path.join(dirname, filename))","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"trainimlist[:3]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Adding image_path to train_im","metadata":{}},{"cell_type":"code","source":"pathlist = []\nfor a in tqdm(train_im[\"id\"]):\n    for b in trainimlist:\n        if a.replace(\"_image\",\"\") in b:\n            pathlist.append(b)\n            break\n            ","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im[\"path\"] = pathlist\ntrain_im.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Show 1 image","metadata":{}},{"cell_type":"code","source":"tmppath = train_im[\"path\"].iloc[0]\ntmppath","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"dataset = pydicom.filereader.dcmread(tmppath)\nimg = dataset.pixel_array","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"plt.imshow(img)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"img.shape","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 2. Analyzing the train data","metadata":{}},{"cell_type":"markdown","source":"#### Count the unique number for each column","metadata":{}},{"cell_type":"code","source":"train_st.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for col in train_st.columns:\n    print(str(col) + \":\" + str(len(train_st[col].unique())))","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"len(train_st)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for col in train_im.columns:\n    print(str(col) + \":\" + str(len(train_im[col].unique())))","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"len(train_im)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### The number of ids in train_st and the number of ids in train_im are not equal\n#### The train_im may have some duplicates in StudyInstanceUID","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## 2.1 Separating from duplicates","metadata":{}},{"cell_type":"code","source":"train_im.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## I think the StudyInstanceUID is like a patient ID","metadata":{}},{"cell_type":"code","source":"uidgroup = train_im.groupby(\"StudyInstanceUID\").count().reset_index()\nuidgroup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"uidgroup = uidgroup.sort_values(\"id\")\nuidgroup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"uidgroup.columns = [\"StudyInstanceUID\",\"id_count\",\"boxes_count\",\"label_count\",\"path_count\"]\nuidgroup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Some of the same StudyInstanceUID(patients) have several images(ids).\n#### Firstly I think about things that are not duplicated(Chapter3). After that, I think about things that are duplicated(Chapter4).","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 3 For no duplicates in StudyInstanceUID\n\n![image.png](https://storage.googleapis.com/kagglesdsdata/datasets/1348626/2246563/chapter3.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=gcp-kaggle-com%40kaggle-161607.iam.gserviceaccount.com%2F20210518%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20210518T151228Z&X-Goog-Expires=259199&X-Goog-SignedHeaders=host&X-Goog-Signature=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)\n","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### extract no duplicates id","metadata":{}},{"cell_type":"code","source":"uidgroup # this dataframe is count table made by groupby function.","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"nodup_count = uidgroup[uidgroup[\"id_count\"]==1] # this dataframe is count table\nnodup_count","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im_no_dup = pd.merge(train_im,nodup_count,on=\"StudyInstanceUID\")\ntrain_im_no_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Clean the column names after extracting only the necessary parts","metadata":{}},{"cell_type":"code","source":"train_im_no_dup = train_im_no_dup[[\"boxes\",\"label\",\"StudyInstanceUID\",\"path\",\"boxes_count\"]]\ntrain_im_no_dup.columns = [\"boxes\",\"label\",\"id\",\"path\",\"box_count\"]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im_no_dup.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Add _study to id to merge with train_st","metadata":{}},{"cell_type":"code","source":"train_im_no_dup[\"id\"] = [s + str(\"_study\") for s in train_im_no_dup[\"id\"]]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im_no_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_st.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_st_no_dup = pd.merge(train_st,train_im_no_dup,on=\"id\")","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_st_no_dup.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Compare in cases where box_count is 0 and 1","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## 3.1 Analyzing for no duplicates by the number of bounding box values.\n\n\n\n\n","metadata":{}},{"cell_type":"markdown","source":"## 3.1.1 Bounding Box with nan","metadata":{}},{"cell_type":"code","source":"bc0 = train_st_no_dup[train_st_no_dup[\"box_count\"]==0]\nbc0","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Since box0 has no bounding boxes and label is none 1 0 0 1 1, I thought it was all \"Negative for Pneumonia\", but different cases were also confirmed.","metadata":{}},{"cell_type":"code","source":"bc0.iloc[:,1:5].sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"bc0.iloc[:,1:5].sum().sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### 81 out of 1705 are Atypical Apperance, 1 is Typical Apperance, and others are Negative for Pneumonia.\n#### You have to think about how to handle this.","metadata":{}},{"cell_type":"markdown","source":"## 3.1.2  Bounding Box with a value","metadata":{}},{"cell_type":"code","source":"bc1 = train_st_no_dup[train_st_no_dup[\"box_count\"]==1]\nbc1","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"bc1.iloc[:,1:5].sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"bc1.iloc[:,1:5].sum().sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Those with a bounding box and label are not \"Negative for Pneumonia\". This is as expected.","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### -----Summary so far------\n\nSome patients have multiple images\n\nIn the case of no duplicates, the bounding box is classified into two types, one with nan and the other without.\n\n    #### 1. case 1 : the bounding box value is nan\n    \n        81 out of 1705 are Atypical Apperance, 1 is Typical Apperance, and others are Negative for Pneumonia.\n        That is, all data is not always \"Negative for Pneumonia\" !   (95% : Negative for Pneumonia, 5 % : other. )\n        \n        You have to think about how to handle this.\n\n        \n    #### 2. case 2 : the bounding box value has a value\n    \n    　　Those with a bounding box and label are not \"Negative for Pneumonia\". This is as expected.\n        \n","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 4. For duplicates in StudyInstanceUID\n\n![image.png](https://storage.googleapis.com/kagglesdsdata/datasets/1348626/2246563/chapter4.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=gcp-kaggle-com%40kaggle-161607.iam.gserviceaccount.com%2F20210518%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20210518T151316Z&X-Goog-Expires=259199&X-Goog-SignedHeaders=host&X-Goog-Signature=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)","metadata":{}},{"cell_type":"markdown","source":"#### Extract duplicates id","metadata":{}},{"cell_type":"code","source":"uidgroup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"uidgroup_dup = uidgroup[uidgroup[\"id_count\"]>=2]\nuidgroup_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"The duplicates id can be extracted.","metadata":{}},{"cell_type":"markdown","source":"## 4.1 Analyzing for duplicates","metadata":{}},{"cell_type":"code","source":"train_im.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"uidgroup_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Showing boxes_count = 0","metadata":{}},{"cell_type":"code","source":"train_im[train_im[\"StudyInstanceUID\"]==uidgroup[\"StudyInstanceUID\"].iloc[-5]]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### All images are labeled by none in this StudyInstanceUID.","metadata":{}},{"cell_type":"markdown","source":"#### Next, showing boxes_count = 0","metadata":{}},{"cell_type":"code","source":"train_im[train_im[\"StudyInstanceUID\"]==uidgroup[\"StudyInstanceUID\"].iloc[-1]]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### In this case, there is a value in one boxex and label among multiple ids","metadata":{}},{"cell_type":"code","source":"train_im[train_im[\"StudyInstanceUID\"]==uidgroup[\"StudyInstanceUID\"].iloc[-2]]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### It is the same as above. If there are multiple ids, is there one box and one label?","metadata":{}},{"cell_type":"code","source":"uidgroup_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"uidgroup_dup.boxes_count.max()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### There is no duplication of boxes, that is, there is only one type bounding box values in multiple ids. \n#### If there is a bounding box in multiple images, try adopting that image","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## 4.2 Adopt and extraction","metadata":{}},{"cell_type":"markdown","source":"#### Rules\n* If Boxes are all nan, drop_duplicate and extract.\n\n\n* If Boxes has a value, the imageid with the value is extracted.","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### In order to extract, firstly merge with count data.","metadata":{}},{"cell_type":"code","source":"train_im","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"uidgroup_dup.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"uidgroup_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im_dup = pd.merge(train_im,uidgroup_dup,on=\"StudyInstanceUID\")\ntrain_im_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Exclude places where box is none and the box count is 1.","metadata":{}},{"cell_type":"markdown","source":"#### Firstly, extract where box is none and the box count is 1.","metadata":{}},{"cell_type":"code","source":"tmp1 = train_im_dup[\"boxes\"].isna()\ntmp2 = train_im_dup[\"boxes_count\"]==1","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"temp_train_im_dup = train_im_dup[tmp1*tmp2]\ntemp_train_im_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Secondly, pull out the others","metadata":{}},{"cell_type":"code","source":"train_im_dup = train_im_dup.drop(temp_train_im_dup.index) ","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im_dup = train_im_dup.reset_index(drop=True)\ntrain_im_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Lastly, drop_duplicates are carried out because boxes are nan and there are still duplicates where boxes_count is 0.\n#### It's like the 2nd to 4th of the above results.","metadata":{}},{"cell_type":"code","source":"train_im_dup = train_im_dup.drop_duplicates([\"boxes\",\"StudyInstanceUID\"])\ntrain_im_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Extraction is done.","metadata":{}},{"cell_type":"markdown","source":"## 4.3 Merge and analysis","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Clean as well as the study for no duplication case(Chapter3)\n\n","metadata":{}},{"cell_type":"code","source":"train_im_dup = train_im_dup[[\"boxes\",\"label\",\"StudyInstanceUID\",\"path\",\"boxes_count\"]]\ntrain_im_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im_dup.columns = [\"boxes\",\"label\",\"id\",\"path\",\"box_count\"]","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### merge with train_st","metadata":{}},{"cell_type":"code","source":"train_st.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_im_dup[\"id\"] = [s + str(\"_study\") for s in train_im_dup[\"id\"]]\ntrain_im_dup.head(3)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_st_dup = pd.merge(train_st,train_im_dup,on=\"id\")\ntrain_st_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## 4.3.1 Analyzing for duplicates and box_count = 0(box is nan)","metadata":{}},{"cell_type":"code","source":"bc0_dup = train_st_dup[train_st_dup[\"box_count\"]==0]\nbc0_dup","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"bc0_dup.iloc[:,1:5].sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"bc0_dup.iloc[:,1:5].sum().sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Although the bounding box value = nan, 2 out of 55 are Atypical Apperance. And others are Negative for Pneumonia.(96.4% : Negative for Pneumonia, 3.6% : other)\n#### You have to think about how to handle this.","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## 4.3.2 Analyzing for duplicates and box_count = 1","metadata":{}},{"cell_type":"code","source":"bc1_dup = train_st_dup[train_st_dup[\"box_count\"]==1]\nbc1_dup\n","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"bc1.iloc[:,1:5].sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"bc1.iloc[:,1:5].sum().sum()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Those with a bounding box and label are not \"Negative for Pneumonia\". This is as expected.","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### -----summary of this part------\n ### Group2. Duplicates of StudyInstanceUID\n   \n   In duplicate ids, the box values is divided into two categories.\n       \n   One is the case of all nan like this.\n           \n  ![image.png](attachment:34e05c27-0b12-4fa8-ba7b-8de8744f9a1a.png)\n           \n           \n           \n       \n   The other is the case that it consists of the ids with one type of bounding box value and nan like this.\n           \n           \n           \n   ![image.png](attachment:ec24a17a-773e-4d97-996d-653a1d4f3575.png)\n           \n           \n           \n       \n   For making clean data and analyzing, the one image id is determined as follows.\n       \n       \n   #### Rules\n   * If Boxes are all nan, drop_duplicate and extract.\n\n\n   * If Boxes has a value, the imageid with the value is extracted.\n   \n       After that, the analysis was performed in the same way as no duplication.\n   \n   \n a) The case that the box is nan in group1.\n\n  2 out of 55 are Atypical Apperance, and others are Negative for Pneumonia.\n  You have to think about how to handle this.\n\n  All are not always Negative for Pneumonia.\n\n b)  The case that there is some value in the box\n\n  All are not \"Negative for Pneumonia\". This is as expected.","metadata":{},"attachments":{"34e05c27-0b12-4fa8-ba7b-8de8744f9a1a.png":{"image/png":"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"},"ec24a17a-773e-4d97-996d-653a1d4f3575.png":{"image/png":"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"}}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 5. Merge All data(Making cleaning data with train_st + train_im)\n\n![image.png](https://storage.googleapis.com/kagglesdsdata/datasets/1348626/2246563/chapter5.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=gcp-kaggle-com%40kaggle-161607.iam.gserviceaccount.com%2F20210518%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20210518T151353Z&X-Goog-Expires=259199&X-Goog-SignedHeaders=host&X-Goog-Signature=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)","metadata":{}},{"cell_type":"code","source":"cleandf = pd.concat([bc0,bc1,bc0_dup,bc1_dup])","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"cleandf","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### confirming whether the number of concated dataframe is same as the train_study_level.csv","metadata":{}},{"cell_type":"code","source":"len(train_st)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### OK","metadata":{}},{"cell_type":"markdown","source":"#### next, I make the id order the same as train_st","metadata":{}},{"cell_type":"code","source":"tmptrain_st = train_st[[\"id\"]]\ntmptrain_st","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"cleandf2 = pd.merge(tmptrain_st,cleandf,on=\"id\",how=\"left\")\ncleandf2","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"#### Save","metadata":{}},{"cell_type":"code","source":"cleandf2.to_csv(\"cleandf.csv\",index=False)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"cleandf2.info()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 6.Summary\n\n## There was duplication of \"StudyInstanceUID\" in the train_image_level.csv.\n## I analyzed for that duplicates and cleanly merge with train_study_level.csv.\n ","metadata":{}},{"cell_type":"markdown","source":"![image.png](https://storage.googleapis.com/kagglesdsdata/datasets/1348626/2246563/Clipboard03.jpg?X-Goog-Algorithm=GOOG4-RSA-SHA256&X-Goog-Credential=gcp-kaggle-com%40kaggle-161607.iam.gserviceaccount.com%2F20210518%2Fauto%2Fstorage%2Fgoog4_request&X-Goog-Date=20210518T151418Z&X-Goog-Expires=259199&X-Goog-SignedHeaders=host&X-Goog-Signature=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)","metadata":{}},{"cell_type":"markdown","source":"# From now on, I will check if this data is correct.\n# Thank you for reading this far. \n# I hope you find it useful, I'm grad to upvoting this notebook !  ","metadata":{}},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}