Photography and profiles
Find animals who may benefit from better photos, a clearer biography, or updated information before their listing becomes stale.
Why SecondChance exists
Help the animals most likely to be overlooked—before they have to wait longer.
Animal shelters make difficult decisions every day with limited time, staff, foster homes, and outreach capacity. SecondChance uses shelter data to help teams notice which animals may need extra attention early, then turn that insight into practical action.
01The problem
Long shelter stays are easier to prevent than to undo.
Some animals are adopted quickly. Others wait because of age, breed perceptions, health or behavior needs, weak photos, limited profile information, seasonal crowding, or simply not reaching the right audience. Staff usually know these patterns, but it is difficult to review every case at the right time while also running the shelter.
SecondChance is designed to support that judgment. It does not replace shelter professionals. It gives them another way to see where earlier attention may make the biggest difference.
02Three concrete uses
Three concrete uses
01
Earlier individual support
An intake-time estimate beside an actual wait, so a specific animal gets a closer look sooner.
02
Shelter-level planning
Intake, outcome and wait trends to anticipate busy periods and spread outreach across the team.
03
Evaluating outreach results
A written plan with its assumptions, so what happened afterwards can be compared against what was expected.
03What the data shows
From individual stays to the shelter's bigger picture.
The Austin Animal Center publishes intake and outcome records. Once those events are cleaned and paired correctly, they can show:
- How many high-confidence shelter stays appear to be open at a selected time.
- How long animals have been waiting.
- Which groups tend to experience longer stays.
- How adoption, transfer, and return-to-owner outcomes change over time.
- When crowding and seasonal intake patterns create more pressure.
- Which animals a validated model estimates may be at higher risk of a long stay.
- Whether predictions remain accurate, calibrated, and fair enough to use.
What this data does not say
A public intake record is not automatically proof that an animal is currently available for adoption. SecondChance separates intake history, inferred current stays, verified adoption listings, and model predictions so users can understand exactly what each number means.
04How a shelter can use it
Turn data into earlier, more focused help.
Foster recruitment
Give foster coordinators a focused list of animals whose stress, age, medical needs, or long wait may make a foster placement especially valuable.
Targeted outreach
Match animals with the channels most likely to help—social media, breed-specific communities, senior-pet audiences, local partners, or adoption events.
Daily queue management
Help teams decide which cases deserve a closer look first without reducing the care given to anyone else.
Medical and behavior support
Surface animals whose barriers may require a clinical, enrichment, training, or behavior-plan review.
Capacity planning
Use intake and outcome trends to anticipate busy periods, prepare kennel space, recruit volunteers, schedule events, and coordinate transfer partners.
Program measurement
Compare what happened after new photos, foster placement, profile rewrites, fee promotions, or events. This helps shelters learn which interventions work, for which animals, and at what cost.
Equity and model monitoring
Check whether the system performs differently across species, ages, conditions, and breed groups. Pause or change the model when it is not reliable enough for a group.
Potential benefits, not measured results
The uses above describe what the workflow makes possible. SecondChance has not measured animals saved, adoption improvements, or outreach effects, and it does not have users or partners to report. Those claims belong to a shelter that runs the workflow and evaluates it.
05Beyond Austin
The workflow can travel. The model cannot be copied blindly.
Shelters across the country can use the same basic approach: combine clean intake and outcome history, define each shelter stay correctly, identify practical intervention points, and measure results. But every shelter has different communities, policies, resources, housing conditions, and data systems.
A model trained on Austin records should not be treated as automatically accurate somewhere else. Each shelter should train or recalibrate with its own data, test the model on a recent time period, review performance by important groups, and keep staff in control of every decision.
Our mission
More support, sooner.
Our mission is to help shelters use data to direct more attention, creativity, and community support toward animals who may otherwise be overlooked. A high-risk score is never a reason to reduce care. It is a signal to ask what more can be done.
06The guardrail
The guardrail is simple: predictions can increase help, never take it away.
- Never use a score to justify euthanasia.
- Never use a score to refuse intake.
- Never use a score to withhold medical or behavioral care.
- Never use a score to lower an animal's priority for food, housing, enrichment, or safety.
- Never present an uncertain prediction as a fact.
- Never hide stale, incomplete, or fallback data behind a “live” label.
The full responsible-use policy
SecondChance estimates how long an animal is likely to wait so that staff attention, foster recruitment, photography and outreach can go to the animals who need them most.
A high predicted long-stay risk is a request for more help, not less. This score must never be used to justify euthanasia, to refuse an intake, to withhold medical or behavioural care, or to deprioritise any animal for any resource. Every prediction in this system exists to increase what an animal receives.
If you are being asked to use this tool to reduce care for an animal, the tool is being misused. Say so.
07Project context
Built to explore how responsible prediction can support animal welfare.
SecondChance began as a UNC Charlotte predictive analytics project using public Austin Animal Center records. The project combines data engineering, predictive modeling, model evaluation, and product design to show how analysis can become a practical, responsible shelter workflow.
Data source: City of Austin Open Data. Current and historical feeds use different schemas because Austin moved to ShelterBuddy in May 2025. SecondChance documents that change and reports the source and freshness of every view.