The 2020 AADT csv at least has been lazily prepared and is consequently unusable. There are commas in text fields, and the text fields are not wrapped in quotation marks, so those intra-field commas are interpreted as field delimiters, shifting field contents one column to the right. Also the hour-interval field has been auto-converted to dates, probably by the file being saved in Excel before upload to the open data portal. You need to have someone who knows what they are doing with csv's prepare them for public release. This reflects very poorly on TMR.
The 2017 and 2018 datasets have duplicate records, i.e. the SITE_ID is not unique within the file. Possibly other years are affected as well, I have not checked all of them as yet. Example in 2018 is site_id 100108 - there are wo records with different AADT values. Which record should be taken as the authoritative count? Noting that within each set of dups, several of the dups have a TDIST value that does NOT fall between TDIST_START and TDIST_END. Can this be used as an indication of an invalid count? Thanks for your assistance.
It is noted that the 2020 kml version of the data, the report links for count sites are available for download. Would it be possible to include report links for 2021 and future releases? It prevents having to request the data all the time from TMR and means we are able to obtain the data we need instantaneously. Really liked the feature in the 2020 version. Thanks
Bad formatting of csv's
The 2020 AADT csv at least has been lazily prepared and is consequently unusable. There are commas in text fields, and the text fields are not wrapped in quotation marks, so those intra-field commas are interpreted as field delimiters, shifting field contents one column to the right. Also the hour-interval field has been auto-converted to dates, probably by the file being saved in Excel before upload to the open data portal. You need to have someone who knows what they are doing with csv's prepare them for public release. This reflects very poorly on TMR.
Duplicate SITE_IDs
The 2017 and 2018 datasets have duplicate records, i.e. the SITE_ID is not unique within the file. Possibly other years are affected as well, I have not checked all of them as yet. Example in 2018 is site_id 100108 - there are wo records with different AADT values. Which record should be taken as the authoritative count? Noting that within each set of dups, several of the dups have a TDIST value that does NOT fall between TDIST_START and TDIST_END. Can this be used as an indication of an invalid count? Thanks for your assistance.
2021 data availability
When will the 2021 traffic census data be published?
Report Links
It is noted that the 2020 kml version of the data, the report links for count sites are available for download. Would it be possible to include report links for 2021 and future releases? It prevents having to request the data all the time from TMR and means we are able to obtain the data we need instantaneously. Really liked the feature in the 2020 version. Thanks