{"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":"# Importing required libraries","metadata":{}},{"cell_type":"code","source":"!pip -q install skimpy\nfrom plotly.offline import init_notebook_mode,iplot\ninit_notebook_mode(connected=True)\nimport numpy as np \nimport pandas as pd \nimport skimpy \nimport plotly.express as px\nimport plotly.offline as py\nimport plotly.graph_objs as go\n","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-03-21T10:50:06.583828Z","iopub.execute_input":"2022-03-21T10:50:06.584405Z","iopub.status.idle":"2022-03-21T10:50:20.101363Z","shell.execute_reply.started":"2022-03-21T10:50:06.584285Z","shell.execute_reply":"2022-03-21T10:50:20.100365Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# What are we supposed to predict \nwe have to predict that the sound contains bird sound or not if the voice note contains sound of bird then target is True otherwise target is false ","metadata":{}},{"cell_type":"markdown","source":"# What is the evaluation metric ? \n## Macro F1 Score. \nlink to understand the evalutation metric: https://medium.com/analytics-vidhya/performance-metrics-for-machine-learning-models-80d7666b432e#:~:text=Macro%20F1%2DScore%3A%20Macro%20F1,Classes%20and%20K%20%E2%88%88%20C.\n\nMacro F1-Score: Macro F1-score (short for macro-averaged F1 score) is used to assess the quality of problems with multiple binary labels or multiple classes.\nMacro F1-score is defined as the average harmonic mean of precision and recall of each class:\nC is the Number of Classes and K ∈ C.\n\n![image.png](attachment:66043906-05be-44e5-9cab-781da1d9186f.png)\n\nMacro F1-score will give the same importance to each label/class. It will be low for models that only perform well on the common classes while performing poorly on the rare classes.","metadata":{},"attachments":{"66043906-05be-44e5-9cab-781da1d9186f.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"# Reading the Data","metadata":{}},{"cell_type":"code","source":"taxo = pd.read_csv(\"../input/birdclef-2022/eBird_Taxonomy_v2021.csv\")\nss = pd.read_csv(\"../input/birdclef-2022/sample_submission.csv\")\ntrain = pd.read_csv(\"../input/birdclef-2022/train_metadata.csv\")\ntest = pd.read_csv(\"../input/birdclef-2022/test.csv\")\nscored = pd.read_json(\"../input/birdclef-2022/scored_birds.json\")\ntrain_meta = pd.read_csv(\"../input/birdclef-2022/train_metadata.csv\")","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:20.102983Z","iopub.execute_input":"2022-03-21T10:50:20.103612Z","iopub.status.idle":"2022-03-21T10:50:20.390542Z","shell.execute_reply.started":"2022-03-21T10:50:20.103568Z","shell.execute_reply":"2022-03-21T10:50:20.389602Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 📷Skimpy","metadata":{}},{"cell_type":"code","source":"skimpy.skim(taxo)","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:20.391742Z","iopub.execute_input":"2022-03-21T10:50:20.392008Z","iopub.status.idle":"2022-03-21T10:50:20.493591Z","shell.execute_reply.started":"2022-03-21T10:50:20.391978Z","shell.execute_reply":"2022-03-21T10:50:20.492782Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"skimpy.skim(train)","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:20.495658Z","iopub.execute_input":"2022-03-21T10:50:20.496059Z","iopub.status.idle":"2022-03-21T10:50:20.583705Z","shell.execute_reply.started":"2022-03-21T10:50:20.496018Z","shell.execute_reply":"2022-03-21T10:50:20.582923Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"skimpy.skim(train_meta)","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:20.585234Z","iopub.execute_input":"2022-03-21T10:50:20.585448Z","iopub.status.idle":"2022-03-21T10:50:20.662780Z","shell.execute_reply.started":"2022-03-21T10:50:20.585423Z","shell.execute_reply":"2022-03-21T10:50:20.661977Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"skimpy.skim(test)","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:20.663914Z","iopub.execute_input":"2022-03-21T10:50:20.664447Z","iopub.status.idle":"2022-03-21T10:50:20.727826Z","shell.execute_reply.started":"2022-03-21T10:50:20.664413Z","shell.execute_reply":"2022-03-21T10:50:20.726897Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# 🖼️ Train ","metadata":{}},{"cell_type":"code","source":"train.head()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:58:46.698471Z","iopub.execute_input":"2022-03-21T10:58:46.698963Z","iopub.status.idle":"2022-03-21T10:58:46.717832Z","shell.execute_reply.started":"2022-03-21T10:58:46.698925Z","shell.execute_reply":"2022-03-21T10:58:46.717070Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig = px.scatter_geo(train ,lat = \"latitude\", lon = \"longitude\", color = \"primary_label\")\nfig.show()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:20.760147Z","iopub.execute_input":"2022-03-21T10:50:20.760340Z","iopub.status.idle":"2022-03-21T10:50:22.406071Z","shell.execute_reply.started":"2022-03-21T10:50:20.760315Z","shell.execute_reply":"2022-03-21T10:50:22.404982Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig = px.bar( x = train.scientific_name.value_counts().keys(), y = train.scientific_name.value_counts().values)\n\nfig.update_layout(\n    title=\"Scientific names frequency\",\n    xaxis_title=\"species name\",\n    yaxis_title=\"species freqency\",\n    font=dict(\n        family=\"Courier New, monospace\",\n        size=10,\n        color=\"RebeccaPurple\"\n    )\n)\nfig.show()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:22.407610Z","iopub.execute_input":"2022-03-21T10:50:22.407930Z","iopub.status.idle":"2022-03-21T10:50:22.513545Z","shell.execute_reply.started":"2022-03-21T10:50:22.407895Z","shell.execute_reply":"2022-03-21T10:50:22.512692Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig = px.bar( x = train.common_name.value_counts().keys(), y = train.common_name.value_counts().values)\n\nfig.update_layout(\n    title=\"common_name frequency\",\n    xaxis_title=\"common_name\",\n    yaxis_title=\"common_name freqency\",\n    font=dict(\n        family=\"Courier New, monospace\",\n        size=10,\n        color=\"RebeccaPurple\"\n    )\n)\nfig.show()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:22.515645Z","iopub.execute_input":"2022-03-21T10:50:22.515865Z","iopub.status.idle":"2022-03-21T10:50:22.594873Z","shell.execute_reply.started":"2022-03-21T10:50:22.515839Z","shell.execute_reply":"2022-03-21T10:50:22.593890Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig = px.bar( x = train.author.value_counts().keys(), y = train.author.value_counts().values)\n\nfig.update_layout(\n    title=\"author frequency\",\n    xaxis_title=\"author\",\n    yaxis_title=\"author freqency\",\n    font=dict(\n        family=\"Courier New, monospace\",\n        size=10,\n        color=\"RebeccaPurple\"\n    )\n)\nfig.show()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:50:22.596023Z","iopub.execute_input":"2022-03-21T10:50:22.596451Z","iopub.status.idle":"2022-03-21T10:50:22.686952Z","shell.execute_reply.started":"2022-03-21T10:50:22.596416Z","shell.execute_reply":"2022-03-21T10:50:22.686064Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig = px.scatter_geo(data_frame = train , lat = \"latitude\", lon =\"longitude\", color = \"rating\", hover_data=[\"rating\", \"primary_label\"])\nfig.update_layout(\n    title=\"rating with primary_labels\",\n    font=dict(\n        family=\"Courier New, monospace\",\n        size=10,\n        color=\"RebeccaPurple\"\n    ),\n    margin=dict(l=40, r=40, t=100, b=80)\n\n)\nfig.show()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:51:13.349626Z","iopub.execute_input":"2022-03-21T10:51:13.349903Z","iopub.status.idle":"2022-03-21T10:51:13.880686Z","shell.execute_reply.started":"2022-03-21T10:51:13.349875Z","shell.execute_reply":"2022-03-21T10:51:13.879801Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig = px.bar(x = train.time.value_counts().keys()[:100], y = train.time.value_counts().values[:100])\nfig.update_layout(\n    title=\"time duration of songs (TOP100)\",\n    font=dict(\n        family=\"Courier New, monospace\",\n        size=10,\n        color=\"RebeccaPurple\"\n    ),\n    xaxis_title=\"time duration of songs\",\n    yaxis_title=\"freqency\",\n    margin=dict(l=40, r=40, t=100, b=80)\n\n)\nfig.show()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T10:55:50.589522Z","iopub.execute_input":"2022-03-21T10:55:50.590169Z","iopub.status.idle":"2022-03-21T10:55:50.670084Z","shell.execute_reply.started":"2022-03-21T10:55:50.590099Z","shell.execute_reply":"2022-03-21T10:55:50.669457Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Taxonomy dataframe vis","metadata":{}},{"cell_type":"code","source":"taxo.head()","metadata":{"execution":{"iopub.status.busy":"2022-03-21T11:10:16.485201Z","iopub.execute_input":"2022-03-21T11:10:16.485520Z","iopub.status.idle":"2022-03-21T11:10:16.500673Z","shell.execute_reply.started":"2022-03-21T11:10:16.485489Z","shell.execute_reply":"2022-03-21T11:10:16.499919Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"fig = px.bar(x = taxo.CATEGORY.value_counts().keys(), y =taxo.CATEGORY.value_counts().values, color = taxo.CATEGORY.value_counts().values ,)\nfig.update_layout(title = \"taxonomy categories\", \n                    font=dict(\n                        family=\"Courier New, monospace\",\n                        size=10,\n                        color=\"RebeccaPurple\"\n                    ),\n                    xaxis_title=\"category\",\n                    yaxis_title=\"no of category\",\n                    margin=dict(l=40, r=40, t=100, b=80)\n                 )","metadata":{"execution":{"iopub.status.busy":"2022-03-21T11:08:37.261472Z","iopub.execute_input":"2022-03-21T11:08:37.262353Z","iopub.status.idle":"2022-03-21T11:08:37.364883Z","shell.execute_reply.started":"2022-03-21T11:08:37.262304Z","shell.execute_reply":"2022-03-21T11:08:37.363949Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}