server issueshttps://gitlab.gistools.geog.uni-heidelberg.de/giscience/missingosm/visualiser/server/-/issues2019-08-01T14:10:48Zhttps://gitlab.gistools.geog.uni-heidelberg.de/giscience/missingosm/visualiser/server/-/issues/8HDBSCAN tweets clustering for whole africa2019-08-01T14:10:48ZHao LiHDBSCAN tweets clustering for whole africaIn order to indentify the OSM missing areas in africa from current tweets data, we should consider the HDBCAN for whole regions and come up with the bounding boxes of tweets clusters, then all the bounding boxes need to be overlapped wit...In order to indentify the OSM missing areas in africa from current tweets data, we should consider the HDBCAN for whole regions and come up with the bounding boxes of tweets clusters, then all the bounding boxes need to be overlapped with current OSM building geometries, and only those does not consist of any valid OSM building geometries would be selected for OSM missing areas.Mohammed ZiaMohammed Ziahttps://gitlab.gistools.geog.uni-heidelberg.de/giscience/missingosm/visualiser/server/-/issues/7from point cluster to prediction extent2019-07-19T16:07:10ZHao Lifrom point cluster to prediction extentFrom the geo-tagged tweets point clusters, we need to further generate the geographical extent of osm missing residential areas (either bounding box or multi-polygons)From the geo-tagged tweets point clusters, we need to further generate the geographical extent of osm missing residential areas (either bounding box or multi-polygons)Hao LiHao Lihttps://gitlab.gistools.geog.uni-heidelberg.de/giscience/missingosm/visualiser/server/-/issues/6HDBSCAN tweets clustering for residential area localization2019-07-19T08:19:37ZHao LiHDBSCAN tweets clustering for residential area localizationFor finding the tweets cluster, we are using the Hierarchical Density-Based Spatial Clustering- HDBSCAN (https://github.com/bobleegogogo/hdbscan). There are two major hyperparameters in HDBSCAN: min_cluster_size and min_samples.
The ide...For finding the tweets cluster, we are using the Hierarchical Density-Based Spatial Clustering- HDBSCAN (https://github.com/bobleegogogo/hdbscan). There are two major hyperparameters in HDBSCAN: min_cluster_size and min_samples.
The ideal case of tweets clusters should be able to catch big cities (high amount of total number) as well as small villages and other types of residential areas (low total number but high density)
Two stage clustering considering different min_cluster_size and min_samples.