{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"_kg_hide-input":true,"_kg_hide-output":true},"cell_type":"code","source":"import numpy as np\nimport pandas as pd\n\nimport matplotlib.pyplot as plt\n%matplotlib inline\n\nimport seaborn as sns\nsns.set()\nsns.set_palette(\"pastel\")\n\nimport warnings\nwarnings.filterwarnings(\"ignore\", category=FutureWarning)","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"c015cff20c28e1f5c7e82a048a9d5d1f75693ec4"},"cell_type":"markdown","source":"Just a first brief look at the data given by this competition and the data that is available on the github page https://github.com/google-research-datasets/gap-coreference.\n\nIn order to directly load the data from github you need to use the raw content of the page and enable the Internet connection in the kernel settings. Probably this is obvious but it took me some time to realize and maybe it helps someone to save time ;)"},{"metadata":{"_uuid":"4ebeab2a0cdc00be7506a87dfe51030e3f7cfae4"},"cell_type":"markdown","source":"## Load GAP Coreference Data\n\n"},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"gap_train = pd.read_csv(\"https://raw.githubusercontent.com/google-research-datasets/gap-coreference/master/gap-development.tsv\", delimiter='\\t')\ngap_test = pd.read_csv(\"https://raw.githubusercontent.com/google-research-datasets/gap-coreference/master/gap-test.tsv\", delimiter='\\t')\ngap_valid = pd.read_csv(\"https://raw.githubusercontent.com/google-research-datasets/gap-coreference/master/gap-validation.tsv\", delimiter='\\t')","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"001ec7244eb860be3df2df6683d453a6200593d7"},"cell_type":"markdown","source":"## Load Competition Data"},{"metadata":{"trusted":true,"_uuid":"68cb3c8c773606946ad213fb303636070a9ddc93"},"cell_type":"code","source":"test_stage_1 = pd.read_csv('../input/test_stage_1.tsv', delimiter='\\t')\nsub = pd.read_csv('../input/sample_submission_stage_1.csv')","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"eab453af692a38b0e6b5956c611b6e78199e9796"},"cell_type":"markdown","source":"## Exploration"},{"metadata":{"_uuid":"634318576a459473a546748f5d1223b10b6aea40"},"cell_type":"markdown","source":"### Gap Train"},{"metadata":{"trusted":true,"_uuid":"8391ac724041caf5761b82676c7f21df581f9f94","_kg_hide-input":true},"cell_type":"code","source":"gap_train.head()","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"fddd34896c86b87a15443531fe6855b320d82ed6"},"cell_type":"markdown","source":"### Gap Test"},{"metadata":{"trusted":true,"_uuid":"5617573eb66b43fc8c1a7825c1f7e363860ea883","_kg_hide-input":true},"cell_type":"code","source":"gap_test.head()","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"93258cee7e47a12e76e7c864321bda1aa847a4cb"},"cell_type":"markdown","source":"### Gap Valid"},{"metadata":{"_kg_hide-input":true,"trusted":true,"_uuid":"b9f4b3624a92b45356adf3111a4fe465fe8c9acd"},"cell_type":"code","source":"gap_valid.head()","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"e478a8cb99c7a10c4c72d05dacb12767e74cd1a7"},"cell_type":"markdown","source":"### Test Stage 1"},{"metadata":{"_kg_hide-input":true,"trusted":true,"_uuid":"dc6b4e18d7cc8e2048c2037c105903651d48ab76"},"cell_type":"code","source":"test_stage_1.head()","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"761f693cdf858a4f047854e245e670783e0409f2"},"cell_type":"markdown","source":"### Countplot Pronouns"},{"metadata":{"trusted":true,"_uuid":"27d79f50ee1ce6c4ba948765c15dc873851c5a37","_kg_hide-input":true},"cell_type":"code","source":"fig, ax = plt.subplots(1, 4, figsize=(20,6))\nordering = ['her', 'his', 'she', 'he', 'She', 'He', 'him', 'Her', 'His', 'hers']\n\nsns.countplot(y='Pronoun',order = ordering, ax=ax[0], data=gap_train)\nax[0].set_title(\"GAP Train\")\n\nsns.countplot(y='Pronoun',order = ordering, ax=ax[1], data=gap_test)\nax[1].set_title(\"GAP Test\")\n\nsns.countplot(y='Pronoun',order = ordering, ax=ax[2], data=gap_valid)\nax[2].set_title(\"GAP Valid\")\n\nsns.countplot(y='Pronoun',order = ordering, ax=ax[3], data=test_stage_1)\nax[3].set_title(\"Test Stage 1\");","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"8a345c8744fffffcefac5c76dd02b932fa303310"},"cell_type":"markdown","source":"### Distplot Pronoun-offset, A-offset, B-offset"},{"metadata":{"_kg_hide-input":true,"trusted":true,"scrolled":false,"_uuid":"89d807c14310cdc2f54104da8da3b26107f691e4"},"cell_type":"code","source":"fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2, 2, figsize=(20,12))\n\nsns.distplot(gap_train[\"Pronoun-offset\"], ax=ax1, label=\"Pronoun-offset\", kde=True)\nsns.distplot(gap_train[\"A-offset\"], ax=ax1, label=\"A-offset\", kde=True)\nsns.distplot(gap_train[\"B-offset\"], ax=ax1, label=\"B-offset\", kde=True)\nax1.set_title(\"GAP Train\")\nax1.set(xlabel='Offset')\n\nsns.distplot(gap_test[\"Pronoun-offset\"], ax=ax2, label=\"Pronoun-offset\", kde=True)\nsns.distplot(gap_test[\"A-offset\"], ax=ax2, label=\"A-offset\", kde=True)\nsns.distplot(gap_test[\"B-offset\"], ax=ax2, label=\"B-offset\", kde=True)\nax2.set_title(\"GAP Test\")\nax2.set(xlabel='Offset')\n\nsns.distplot(gap_valid[\"Pronoun-offset\"], ax=ax4, label=\"Pronoun-offset\", kde=True)\nsns.distplot(gap_valid[\"A-offset\"], ax=ax4, label=\"A-offset\", kde=True)\nsns.distplot(gap_valid[\"B-offset\"], ax=ax4, label=\"B-offset\", kde=True)\nax4.set_title(\"GAP Valid\")\nax4.set(xlabel='Offset')\n\nsns.distplot(test_stage_1[\"Pronoun-offset\"], ax=ax3, label=\"Pronoun-offset\", kde=True)\nsns.distplot(test_stage_1[\"A-offset\"], ax=ax3, label=\"A-offset\", kde=True)\nsns.distplot(test_stage_1[\"B-offset\"], ax=ax3, label=\"B-offset\", kde=True)\nax3.set_title(\"Test Stage 1\")\nax3.set(xlabel='Offset')\nplt.legend();","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"295e58846b474db10a3b1bf8b5c02a4152721ee2"},"cell_type":"markdown","source":"### Countplot A-coref"},{"metadata":{"trusted":true,"_uuid":"99977a8962e73f3d125995d9780b02c0c6b33dbd","_kg_hide-input":true},"cell_type":"code","source":"fig, ax = plt.subplots(1, 3, figsize=(20,3))\nordering = [True, False]\n\nsns.countplot(y='A-coref', order = ordering, ax=ax[0], data=gap_train)\nax[0].set_title(\"GAP Train\")\n\nsns.countplot(y='A-coref',order = ordering, ax=ax[1], data=gap_test)\nax[1].set_title(\"GAP Test\")\n\nsns.countplot(y='A-coref',order = ordering, ax=ax[2], data=gap_valid)\nax[2].set_title(\"GAP Valid\");","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"2484aafea36d797844f021e50474c677c957a1e0"},"cell_type":"markdown","source":"### Countplot B-coref"},{"metadata":{"trusted":true,"_uuid":"b05e2bf831e5d7bd021011c2a0d4ad8e87218cf8","_kg_hide-input":true},"cell_type":"code","source":"fig, ax = plt.subplots(1, 3, figsize=(20,3))\nordering = [True, False]\n\nsns.countplot(y='B-coref', order = ordering, ax=ax[0], data=gap_train)\nax[0].set_title(\"GAP Train\")\n\nsns.countplot(y='B-coref',order = ordering, ax=ax[1], data=gap_test)\nax[1].set_title(\"GAP Test\")\n\nsns.countplot(y='B-coref',order = ordering, ax=ax[2], data=gap_valid)\nax[2].set_title(\"GAP Valid\");","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.6","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}