Data from: Maps made with smartphones highlight lower noise pollution during COVID-19 pandemic lockdown at four locations in Boston

# Data from: Maps made with smartphones highlight lower noise pollution during COVID-19 pandemic lockdown at four locations in Boston [https://doi.org/10.5061/dryad.ncjsxkt35](https://doi.org/10.5061/dryad.ncjsxkt35) Dataset contents include csv files of all data (each file describes collection year and site of data), R script used to create noise maps, and kml files needed to run the map creation code. ## Description of the data and file structure Each csv file contains the L50 values (median sound level) taken from hundreds of 20 second recordings over multiple collection days. The SPLnFFT application exports the latitude and longitude of where the recording was taken, which is also included in the csv files and is used to create the noise maps. The csv files are used as data frames for the R script to create noise maps for each collection site. The R script contains comments and instructions to clearly indicate each step of the map creation. The kml files are used to create boundaries/outlines of the noise maps from Google Maps. The csv and kml files should be kept in the working directory in order to run the code properly. ## Code/Software We used the “ggmap” package in R Studio to overlay the decibel readings over the Google maps of each site. To compare sound levels between different time periods, we created base grids for the four sites that covered the area of data collection using R packages “sf,” “units,” “ggplot2,” “grid,” and “ggsn”. Rather than using raw data within the grid, we used kriging interpolation to standardize the comparison, avoid bias from collection patterns, and create accurate sound maps. We used the R package “gstat” along with “sp,” “ggplot2,” and “ggmap” to create the comparison maps that demonstrated the difference between the kriging interpolations for the two years. The R script contains comments and explanations at each step for clarity.

# 数据源：使用智能手机绘制的地图显示，波士顿四处地点在新冠疫情（COVID-19 pandemic）封锁期间的噪声污染水平更低 [https://doi.org/10.5061/dryad.ncjsxkt35](https://doi.org/10.5061/dryad.ncjsxkt35)
本数据集包含全部实验数据的CSV文件（每个文件标注了数据采集年份与采集站点）、用于绘制噪声地图的R脚本，以及运行地图绘制代码所需的KML（Keyhole Markup Language）文件。

## 数据与文件结构说明
每个CSV文件均包含多日采集的数百条20秒录音对应的L50值（等效声级中值，median sound level）。SPLnFFT应用程序可导出录音采集点的经纬度信息，该信息同样包含于CSV文件中，用于绘制噪声地图。CSV文件将作为数据框供R脚本使用，以绘制各采集站点的噪声地图。该R脚本内置注释与操作说明，清晰标注了地图绘制的每一个步骤。KML文件用于从谷歌地图（Google Maps）中提取噪声地图的边界轮廓。为确保代码正常运行，需将CSV与KML文件置于工作目录下。

## 代码与软件
本研究使用R Studio中的"ggmap"工具包，将分贝读数叠加至各站点的谷歌地图之上。为实现不同时段声级的对比分析，我们借助R语言的"sf"、"units"、"ggplot2"、"grid"与"ggsn"工具包，为覆盖数据采集范围的四个站点生成基础网格。相较于直接使用网格内的原始数据，我们采用克里金插值（Kriging Interpolation）方法以标准化对比流程，规避采集模式带来的偏差，并生成高精度的声级地图。我们结合使用R语言的"gstat"、"sp"、"ggplot2"与"ggmap"工具包，生成对比地图以展示两个年度的克里金插值结果差异。该R脚本的每一步骤均配有注释与解释说明，便于理解与复现。