Predicting upcoming missed clockings and alerting workers or managers
Workers often forget to enter clockings indicating when they have arrived or left work or transferred between work activities. Various embodiments provide a system that predicts the risk of potential missed worker clockings and delivers alerts to the workers and/or their managers when that risk is high. The system analyzes historical data to find previous instances of missed clockings at an organization, extracts feature information regarding the context of the missed clockings, and calibrates a missing clocking risk model. The model is then used in either an online or batch mode to predict the risk of workers missing upcoming scheduled clockings. Based on these risk scores and system configurations, alerts are generated for employees and managers reminding them about the upcoming clockings and lowering the amount of missed clockings and subsequent remedies.
1. A method for providing alerts prior to potential missed clockings, the method comprising:
retrieving, via at least one processor, historical employee clockings data;
identifying, via the at least one processor, missed clockings based on an alignment of the historical employee clockings data;
extracting, via the at least one processor, labels each of which identifies one of the missed clockings as having been missed;
extracting, via the at least one processor, features representative of the missed clockings;
generating, via the at least one processor, a machine-learning-based predictive model trained to generate a risk score based at least in part on the labels and statistical correlations of the extracted features, the risk score indicative of a likelihood that a scheduled clocking will be missed;
calculating, via the at least one processor, the extracted features for an upcoming scheduled clocking within a prediction window;
generating, via the at least one processor using the machine-learning-based predictive model and based at least in part on the calculated extracted features, a risk score for the upcoming scheduled clocking;
determining, via the at least one processor and based at least in part on the risk score, that an alert corresponding to the upcoming scheduled clocking should be presented; and
presenting, via an electronic device, the alert prior to an occurrence of the upcoming scheduled clocking;
wherein:
generating a risk score for the upcoming scheduled clocking is performed in response to at least one of: a schedule being posted, a schedule being updated, or a missed clocking,
extracting features representative of the missed clockings comprises identifying an employee's miss rate for at least one clocking based at least in part on a location, and
the alert includes a response option that provides for the employee to clock in.
2. The method of claim 1 , wherein identifying missed clockings further comprises:
comparing scheduled clockings from a historical employee schedule against actual clockings from the historical employee clockings data in order to identify missed clockings.
3. The method of claim 2 , wherein a scheduled clocking with no associated actual clocking within an allowable buffer for matching is identified as a missed clocking.
4. The method of claim 1 , wherein generating a risk score for the upcoming scheduled clocking comprises identifying each upcoming and scoring scheduled clocking in a given timeframe.
5. The method of claim 4 , wherein the timeframe represents the next day.
6. The method of claim 1 , wherein presenting an alert further comprises:
determining whether the risk score meets a given threshold for an associated user; and
in response to the risk score meeting the given threshold, alerting the associated user prior to the upcoming scheduled clocking.
7. The method of claim 6 , wherein the upcoming scheduled clocking is associated with the employee, and the associated user is the employee.
8. The method of claim 1 , wherein presenting the alert comprises repeating the alert unless a user acknowledgement is received.
9. The method of claim 1 , wherein presenting the alert is performed in accordance with alert configurations.
10. The method of claim 1 , wherein extracting features representative of the missed clockings further comprises:
identifying an employee's miss rate for at least one of:
clockings within a recent time period,
all historical clockings,
clockings at a given time of day, or
clockings on a given day of the week.
11. The method of claim 1 , wherein extracting features representative of the missed clockings is performed in parallel on multiple servers.
12. An apparatus for providing alerts prior to potential missed clockings, the apparatus comprising:
at least one processor; and
at least one memory including computer program code, the at least one memory and the computer program code configured to, with the at least one processor, cause the apparatus to perform at least the following:
retrieve historical employee clockings data;
identify missed clockings based on an alignment of the historical employee clockings data;
extract labels each of which identifies one of the missed clockings as having been missed;
extract features representative of the missed clockings via identifying an employee's miss rate for at least one clocking based at least in part on a location;
generate a machine-learning-based predictive model trained to generate a risk score based at least in part on the labels and statistical correlations of the extracted features, the risk score indicative of a likelihood that a scheduled clocking will be missed;
calculate the extracted features for an upcoming scheduled clocking within a prediction window;
generate, using the machine-learning-based predictive model, a risk score for the upcoming scheduled clocking based at least in part on the calculated extracted features, and in response to at least one of: a schedule being posted a schedule being updated, or a missed clocking;
determine, based at least in part on the risk score, that an alert corresponding to the upcoming scheduled clocking should be presented; and
present, via an electronic device, the alert prior to an occurrence of the upcoming scheduled clocking, wherein the alert includes a response option that provides for the employee to clock in.
13. The apparatus of claim 12 , wherein presenting an alert comprises:
determining whether the risk score meets a given threshold for an associated user; and
in response to the risk score meeting the given threshold, alerting the associated user prior to the upcoming scheduled clocking.
14. The apparatus of claim 12 , wherein the apparatus comprises:
a plurality of processors;
wherein extracting features representative of the missed clockings is performed in parallel on the plurality of processors.
15. A non-transitory computer-readable medium tangibly encoded with a computer program for providing alerts prior to potential missed clockings, the computer program executable by a processor to perform actions comprising:
retrieving historical employee clockings data;
identifying missed clockings based on an alignment of the historical employee clockings data;
extracting labels each of which identifies one of the missed clockings as having been missed;
extracting features representative of the missed clockings;
generating a machine-learning-based predictive model trained to generate a risk score based at least in part on the labels and statistical correlations of the extracted features, the risk score indicative of a likelihood that a scheduled clocking will be missed;
calculating the extracted features for an upcoming scheduled cocking within a prediction window;
generating, using the machine-learning-based predictive model and based at least in part on the calculated extracted features, a risk score for the upcoming scheduled clocking, wherein generating the risk score is in response to at least one of a schedule being posted, a schedule being updated, or a missed clocking;
determining, based at least in part on the risk score, that an alert corresponding to the upcoming scheduled clocking should be presented; and
presenting, via an electronic device, the alert prior to an occurrence of the upcoming scheduled clocking;
wherein the alert includes a response option that provides for an employee to clock in.
16. The non-transitory computer-readable medium of claim 15 , wherein presenting an alert comprises:
determining whether the risk score meets a given threshold for an associated user; and
in response to the risk score meeting the given threshold, alerting the associated user prior to the upcoming scheduled clocking.
17. The method of claim 1 , wherein generating the machine-learning-based predictive model is based at least in part on a Gradient Boosted Decision Tree.
18. The method of claim 17 , wherein the machine-learning-based predictive model is structured to receive the calculated extracted features as a vector.
19. The method of claim 18 , wherein:
the employee clockings data corresponds to an employee; and
the calculated extracted features include an average rate missed clockings rate of the employee.
20. The method of claim 19 , wherein the extracted features further include an average rate of missed clockings for an organization level corresponding to an employee.
21. The method of claim 20 , wherein the extracted features further include a specific miss rate for a day of the week.
22. The method of claim 21 , wherein the alert includes the risk score.
23. The method of claim 22 , wherein the prediction window corresponds to a duration of about a week that occurs in the future.