that occurs to each undertaking supervisor: the stand-up assembly begins and instantly a important ticket is blocked, a developer calls in sick, a dependency slips or a key function is delayed. Immediately, your fastidiously deliberate timeline begins to break down, and also you’re scrambling for options.
On this article, we’ll discover how a machine studying mannequin predicted 41% of undertaking delays earlier than they hit the timeline, reducing prices and lowering last-minute firefighting.
The issue: 62% of IT tasks miss their deadlines in 2025
As a Venture Supervisor working with Agile groups, I’ve usually handled delays and blockers, they rapidly grew to become a part of on a regular basis life. However once I got here throughout the 2025 Wellington State of Venture Administration research revealing that in 2025, 62% of IT tasks miss their deadlines, it shocked me into motion. It’s a rise in comparison with the 2017 PMI Pulse of the Career research the place it was 51% in 2017. Venture delays are reaching a important stage.
I do know delays are frequent, however I hadn’t imagined it might be that prime. However at this time, now we have instruments to anticipate and higher perceive these dangers. Utilizing Python and Information Science, I constructed a mannequin to foretell undertaking delays earlier than they occurred.
This statistic highlights two important factors: delays usually stem from recurring causes, and so they carry an enormous enterprise impression. On this article, we’ll discover how data-driven approaches can uncover these causes and assist undertaking managers anticipate them.
With this data, we will select the most effective plan of action.
That is the place we will use knowledge science. Surprisingly, the 2020 Wellington State of Venture Administration report proclaims that solely 23% of firms use undertaking administration software program to handle their tasks, although these instruments generate a wealth of beneficial knowledge.
By analyzing data from undertaking tickets, we will construct predictive machine studying fashions that spotlight potential dangers earlier than they escalate.
That’s precisely what I did: I analyzed greater than 5,000 tickets, not solely from my present undertaking but additionally from previous initiatives.
Venture administration software program, it seems, is an unimaginable supply of knowledge ready to be leveraged.
The Information Hole in Venture Administration
In conventional undertaking administration, reporting performs a central function, but few reviews provide a complete, detailed retrospective of the undertaking as a complete.
For instance, in Scrum, we observe our velocity, observe the development of our burndown chart, and measure the variety of story factors accomplished.
Conventional reporting nonetheless fails to present us the entire image. Information science can.
As Venture Managers, we might know from expertise the place the important factors lie, however validating these assumptions with knowledge makes our choices much more dependable.
Constructing the Dataset
To discover this concept, I analyzed 5,000 Jira tickets—one of many richest sources of undertaking knowledge out there.
Since actual undertaking knowledge can’t at all times be shared, I generated an artificial dataset in Python that mirrors actuality, together with key variables resembling precedence, story factors, staff dimension, dependencies, and delay.
Code by writer
Having constructed a sensible dataset, we will now discover the totally different ticket profiles it incorporates. This units the stage for our exploratory knowledge evaluation.
Most tickets are of low or medium precedence, which is according to how undertaking backlogs are normally structured. This preliminary distribution already hints at the place dangers may accumulate, a degree we’ll discover additional within the EDA.
Whereas excessive and important precedence tickets characterize a smaller share of the whole, they’re disproportionately extra more likely to be delayed.
This bar plot confirms the phenomenon: high-priority tickets are strongly related to delays. Nevertheless, this could stem from two totally different dynamics:
- Excessive-priority tickets are inherently extra advanced and subsequently at larger threat of delay.
- Some tickets solely develop into excessive precedence as a result of they had been delayed within the first place, making a vicious cycle of escalation.
With this simulated dataset, we now have a sensible snapshot of what occurs in actual tasks: tickets differ in dimension, dependencies, and complexity and a few inevitably find yourself delayed. This displays the on a regular basis challenges undertaking managers face.
The following step is to maneuver past easy counts and uncover the patterns hidden within the knowledge. By means of Exploratory Information Evaluation (EDA), we will check our assumptions: do greater priorities and extra dependencies actually improve the chance of delay? Let’s discover out.
Exploratory Information Evaluation (EDA)
Earlier than shifting to modeling, it’s necessary to step again and visualize how our variables work together. Exploratory Information Evaluation (EDA) permits us to uncover patterns in:
- How delays differ with precedence.
- The impression of dependencies.
- The distribution of story factors.
- The everyday staff sizes dealing with the tickets.
This chart confirms a key instinct: the upper the precedence, the larger the likelihood of delay.
Dependencies amplify this impact, the extra there are, the upper the possibilities of one thing slipping by.
As soon as a delay or threat of delay seems, escalation mechanisms push the precedence even greater, making a suggestions loop.
Lastly, ticket complexity additionally performs a task, including one other layer of uncertainty.
Most tickets fall into the medium-risk class. These demand essentially the most consideration from undertaking managers: whereas not important at first, their threat can rapidly escalate and set off delays.
Excessive-risk tickets, although fewer, carry a disproportionate impression if not managed early.
In the meantime, low-risk tickets normally require lighter monitoring, permitting managers to focus their time the place it really issues.
We additionally discover that the majority tickets have small story level sizes, and groups are normally round 5 members.
This means that agile practices are usually being adopted.
Now, we are going to go additional and take a look at the distribution of threat scores throughout tickets.
We see that solely a small portion of tickets carry a really excessive threat, whereas most sit within the medium zone. Which means by focusing early on the riskiest tickets, Venture Managers may stop many delays.
To check this assumption, let’s now discover how complexity per individual and precedence work together with threat scores.
We can not observe a transparent development right here. The chance rating doesn’t appear to strongly rely on both ticket complexity or precedence, suggesting that different hidden components may drive delays.
Technical Deep Dive: Predictive Mannequin
The uncooked knowledge gives a stable basis, however area data is crucial to constructing a very strong mannequin. To higher seize the dynamics of real-world tasks, we engineered new options that replicate undertaking administration realities:
- Complexity per individual = story factors/staff dimension.
- Has dependency = whether or not a ticket will depend on others (dependencies > 0).
- Precedence story factors interplay = precedence stage multiplied by story factors.
Code by writer
We selected a Random Forest mannequin as a result of it might probably deal with non-linear relationships and gives insights into function significance.
Our predominant focus is on Recall for the optimistic class (1 = delayed). As an illustration, a recall of 0.6 would imply the mannequin appropriately identifies 60% of all really delayed tickets.
The target isn’t good precision however early detection. In undertaking administration, it’s higher to flag potential delays, even with some false positives, than to overlook important points that would derail your entire undertaking.
Code by writer
The mannequin achieved a recall of 0.41, that means it efficiently detected 41% of the delayed tickets.
This will appear modest. Nevertheless, in a undertaking administration context, even this stage of early warning is effective. It offers Venture Managers actionable indicators to anticipate dangers and put together mitigations.
With additional refinement, the mannequin may be improved to anticipate extra delays and assist stop points earlier than they materialize.
We’ll use a confusion matrix to raised perceive the mannequin’s strengths and weaknesses.
Code by writer
The mannequin appropriately identifies 169 delays, nevertheless it additionally generates 373 false alarms, duties flagged as delayed that really completed on time. For a Venture Supervisor, this trade-off is suitable as a result of it’s higher to research a number of false positives than to overlook a important delay. This is part of threat administration.
Nevertheless, the mannequin nonetheless misses 245 delayed tickets, that means its predictions are removed from good.
Total, this mannequin works greatest as an early warning system. It gives beneficial indicators however nonetheless wants additional coaching and refinement. Most significantly, it ought to be complemented with human experience, the judgment and expertise of Venture Managers, to make sure a whole and dependable undertaking overview.
Mannequin Interpretability, Scoring, Enterprise impression, Dashboard & mannequin validation
To actually perceive why the mannequin makes these predictions, we have to look underneath the hood. Which options drive the chance of delay essentially the most? That is the place mannequin interpretability is available in.
Code by writer
We will observe that complexity and the priority-story factors interplay are the strongest drivers of prediction accuracy.
Scoring tickets: Figuring out What’s Actually at Threat.
Why does this matter for Venture Managers? As a result of we will go one step additional.
Calculate a threat rating for every ticket.
This rating highlights which duties are most in danger, permitting PMs to focus their consideration the place it issues most and take preventive motion earlier than delays escalate.
Code by writer
Enterprise Affect Evaluation.
The tickets with the best threat scores affirm the development: solely excessive and important precedence duties carry the best threat.
This perception issues not just for managing the undertaking timeline but additionally for its monetary impression on the enterprise. Delays don’t simply decelerate supply, they improve prices, scale back shopper satisfaction, and eat beneficial staff sources.
To quantify this, we will estimate the enterprise worth of the predictions by simulating how a lot value may be averted once we anticipate dangers and take preventive motion.
Code by writer
Our baseline exhibits that 27.6% of tickets are delayed. However what if Venture Managers may focus solely on the riskiest 20%? We’ll now simulate this focused intervention and see how a lot impression it makes.
Code by writer
We recognized 1,021 high-risk tickets, representing about 20% of all duties. Amongst them, 516 (50.5%) are literally delayed. In different phrases, these few tickets alone drive roughly 10% of complete undertaking delays.
To make this extra concrete, we translate the impression into enterprise phrases with a medium-size undertaking valued at $100,000. By making use of preventive actions on these high-risk tickets, we will estimate the potential value financial savings.
Code by writer
By taking early actions, we may save $9,270, almost 10% of the whole undertaking value. That’s not simply threat mitigation; it’s a direct enterprise benefit.
PM Dashboard
To make these insights actionable, we will additionally construct a Venture Administration Dashboard. It offers a real-time view of dash well being, with all the important thing KPIs wanted to trace progress, anticipate dangers, and preserve a whole undertaking overview.
Code by writer
Mannequin validation
We examined the robustness of the mannequin with a 5-fold cross-validation. Recall was chosen as the primary metric, as a result of in undertaking administration it’s extra necessary to catch potential delays than to maximise total accuracy.
Code by writer
The recall scores throughout folds ranged from 0.39 to 0.42. This implies the mannequin is much from flawless, nevertheless it persistently flags round 40% of delays, a beneficial early warning that helps undertaking managers act earlier than points escalate.
Conclusion
In conclusion, this text confirmed how knowledge science may help make tasks smoother by offering a clearer understanding of the causes of delays.
Information doesn’t exchange the instinct of a Venture Supervisor, nevertheless it strengthens it, very like giving a pilot higher devices to navigate with precision and with a greater view of what’s occurring.
By predicting dangers and figuring out at-risk tickets, we will scale back delays, stop conflicts, and finally ship extra worth.
Venture Managers ought to embrace knowledge science. As we speak, there are two kinds of PMs: conventional ones and data-driven ones. They don’t compete in the identical league.
Lastly, these abilities will not be restricted to undertaking administration. They prolong to product administration and enterprise evaluation. Studying SQL or Python enhances your means to collaborate with builders, perceive product efficiency, and talk successfully throughout all ranges of the enterprise.
Lesson for Venture Managers
What number of of our undertaking choices are based mostly on so-called “greatest practices” which are, in truth, unverified assumptions? Whether or not it’s about assembly schedules, staff construction, or communication strategies, knowledge may help us problem our biases and uncover what really works.
Relying on the group, the evaluation can even go deeper: grouping tickets by undertaking part, subject, or stakeholder might reveal hidden bottlenecks and systemic points.
For instance, velocity usually drops through the QA part. Is it as a result of QA engineers underperform? Under no circumstances. They do their job very effectively. The actual concern is the fixed back-and-forth with builders: clarifying tickets, determining how testing ought to be carried out, or asking for lacking data.
To unravel this, we launched a easy course of: builders now add clearer testing particulars within the ticket and spend 5 minutes on a fast handover name with QA. That small funding of time boosted staff productiveness and velocity by greater than 15%.
Who am I?
I’m Yassin, an IT Venture Supervisor who determined to study Information Science, Python, and SQL to bridge the hole between enterprise wants and technical options. This journey has taught me that essentially the most beneficial undertaking insights come from combining area experience with data-driven approaches. Let’s join on LinkedIn
