Monitor disaster signals in social media module instructor materials

In this module, students explore how social media posts become geographic data during a disaster, using simulated posts about the February 2023 Türkiye–Syria earthquakes. Through a guided storymap and a hands-on lab, students classify posts by intent using a text classification model, extract place entities from unstructured text with gazetteer and deep learning (NER) approaches, geocode extracted place names, and summarize post locations at the province level. Students then compare post-derived location counts with recorded casualties to evaluate whether social media intensity tracks disaster impacts, and conclude by identifying and communicating the uncertainty introduced at each step of the posts-to-map pipeline — confronting the ethical risks of turning vague text into precise-looking maps. Technical Requirements: This module uses ArcGIS Pro and requires students to have an ArcGIS Pro Advanced license with the Deep Learning Libraries installed. GISCI:603, GISCI:405, GISCI:501, GISCI:605, GISCI:205本模块中，学员将依托针对2023年2月土耳其-叙利亚地震的模拟社交媒体帖文，探究灾害场景下社交媒体帖文向地理数据的转化路径。通过引导式故事地图与实操实验单元，学员将借助文本分类模型按意图对帖文进行分类，结合地名词典（gazetteer）与深度学习（命名实体识别，Named Entity Recognition，NER）方法从非结构化文本中提取地点实体，对提取得到的地名开展地理编码，并按省级尺度汇总帖文的发布位置。随后学员将对比基于帖文提取的位置统计量与记录的伤亡数据，以评估社交媒体活跃度是否与灾害影响程度相契合；最后学员将梳理并阐释帖文转制图全流程各环节引入的不确定性，直面将模糊文本转化为看似精准的地图所蕴含的伦理风险。技术要求：本模块需使用ArcGIS Pro软件，学员需持有安装了深度学习库的ArcGIS Pro高级授权许可。课程编号：GISCI:603、GISCI:405、GISCI:501、GISCI:605、GISCI:205