There are two methods you can use to upload data for predictions:
SFTP is the preferred, suggested method. While some level of effort is required to set this up, there's little to no future effort involved, as evidenced in this comparison:
Criteria |
SFTP |
In-App |
---|---|---|
Setup Effort |
Medium |
Low |
Recurring Effort |
None |
High |
Automated |
Yes |
No |
Othot Data Extract |
Automated |
Manual |
Quality |
No Error |
Possible Human Error |
Data Benefits |
Up-to-date |
Status |
Regardless of the data upload option you choose to use, the files must be in .csv format and abide by the IETF RFC 4180 standard. Additionally, we recommend you use the following file naming convention:
<customer>_<division>_<hiq>_<file source>_<train/predict>_<date>.csv |
<customer> |
Your institution's name. An abbreviation is acceptable (similar to a domain name). |
<division> |
Required if there are separate subgroups or instances under the primary customer; this may be excluded if not applicable. |
<hiq> |
description of a high impact question (HIQ) and applicable in case multiple HIQs exist now or in the future; an example for the likelihood for enrollment might be <enroll>. |
<file source> |
primary description for where the file comes from or what the file contains; might identify a CRM or SIS file or a description of what data is coming from the CRM or SIS. |
<train/predict> |
train designation is for historic files with known outcomes that will be used to build a predictive model; predict designation is for current files that will be used to predict on a current or future population. |
<date> |
the date the current data file is extracted or delivered from the source application; required for differentiation between expected daily or weekly predict files |
To set up this method, we'll need to work with a qualified technical contact at your institution. Once you've identified this individual, have them review the following information.
Please send us the following information:
FileZilla is a GUI application that allows you to connect to remote servers. You can access our SFTP as follows:
We recommend the following settings for your export from Slate. For more information on this process, review their knowledge base article.
To keep your predictions up-to-date in the platform, we need to make sure that the right files are delivered and that they are ready to be processed. When using SFTP, please:
Using the In-App Method
If you select this method, use the DATA page to upload your files. Remember to adhere to the file formatting and naming conventions noted above.