Abstract
We collected data from visitors to two urban lakes in Saint Paul, Minnesota, using a conversational chatbot to assess visitor perception of current lake water quality, trends in water quality over time, and other questions relevant to park managers. Data were collected at Como Lake in 2022, 2023, and 2024 and at Lake Phalen in 2023 and 2024. Signs were installed at three locations around each lake with high pedestrian traffic. Each sign had a hook question (“How many watercraft are on the lake right now? Text the number to XXX-XXX-XXXX”). Visitors who responded to this question initiated a series of optional follow-up questions, using a conversational chatbot run by software that automates the sending and receiving of text messages. Survey questions included asking respondents about their primary purpose for visiting the lake today, how often they have visited the lake in the past 12 months, and whether they perceive that water quality in the lake is improving, getting worse, or remaining about the same. Respondents were asked to provide their ZIP code, used to estimate distance traveled to the lake. Respondents also had the opportunity to opt-in to future data collection via phone or text. Unique anonymous identifiers were used to key survey response data. Some responses were parsed from the raw text for numerical values, broader categories, or positive or negative sentiment. These responses are under columns with the "body clean_" label at the beginning of the header in the datasets.