Student Life

What students do outside class

Which activities students choose, how that is shifting, which organizations they join, stay in and lead, and where it leads. From the team's student-experience notebook.

Activities

In [2]:pctCur.sort_values().plot.barh()

Student organizations are the most common experience (66% of current students), then internships (44%) and campus jobs (42%). Co-ops are the rarest at 8%.

In [3]:stuExp.groupby("campus_id")["experience_type"].nunique()

Most students try 2 to 4 different types of experience. 181 people have tried none, and nobody has tried all ten.

In [4]:typeShare["2024–26"] - typeShare["2018–20"]

Certifications gained the most ground (+1.8 pts) and internships lost the most (−4.3 pts). These are shares, not raw counts, because current students make recent years busier for every type.

In [5]:shareByYear = counts.div(counts.sum(axis=1), axis=0)

Internships peaked at 27% of all participation in 2021 and fell to 17% by 2026, while certifications jumped to 21% and overtook them.

Organizations

In [6]:orgPeople.nlargest(10, "people")

The Division of Information Technology campus job reaches the most people (561). Six of the top 10 are hackathons.

In [7]:share.nlargest(10, "change")

Brightline Code School grew the most (+0.49 pts of all participation), with two cloud groups and a security lab close behind.

In [8]:(clubs["outcome"] != "Left after one term").groupby(clubs["experience_name"]).mean()

Retention is similar everywhere: the best clubs keep 80% of members past the first term, barely above the 75% average for all clubs.

In [9]:isLeader.groupby(org[groups]).mean().nlargest(10)

In the Collegiate Cyber Defense Team, 28% of members held a leadership role, against 19% across all clubs and teams.

In [10]:firstGenRate[firstGenRate["members"] >= 50].nlargest(10, "pct")

The National Society of Black Engineers chapter has the highest share of first-generation members (41% against 28% campus-wide). For 7 of these 10 organizations the interval crosses the campus-wide line, so their difference could be chance.

Where it leads

In [11]:offerRate[offerRate["size"] >= 30].nlargest(10, "mean")

Copper Ridge Data Centers turns 52% of internships and co-ops into return offers, against 41% across all companies.

In [12]:orgRank[orgRank["n"] >= 30].nlargest(10, "premium")

Three cybersecurity certification bodies lead, each about $10K above similar grads, followed by consulting and software internship employers.

In [13]:ax.boxplot([adj[grp == k] for k in range(11)])

The median first-job salary rises with the number of experience types, from about $12K below similar grads with none to about $7K above with six. The groups with 8 or more are tiny (5, 1 and 0 alumni).

In [14]:entry.pivot(index="job_family", columns="year", values="median_k")

Cybersecurity pays the most at entry level in almost every year (about $89K on average), and IT Support & Operations the least (about $60K).

Smart Moves

Which in-school choices pay off in job hunting

Alumni only: they have both campus records and job outcomes. Salary premium = first-job salary minus the median of grads with the same major and graduation year. From the team's strategies notebook.

Internships and job routes

In [15]:a.plot_time_to_hire(intern["df_intern"])

Most employed alumni with 2 or 3 internships were hired within a month of graduation, while those with 0 or 1 took much longer. A 3rd internship adds little over a 2nd for both salary and hire time.

In [16]:stuAlum.groupby("first_job_found_via")["sal_adj"].median()

Only return offers pay above similar grads (+$7.2K median). Every other route lands $1K to $4K below.

In [17]:offerRate[offerRate["size"] >= 30].nlargest(10, "mean")

Copper Ridge Data Centers turns 52% of internships and co-ops into return offers, against 41% across all companies.

What to do on campus

In [18]:stats.mannwhitneyu(withSalary.loc[did, "sal_adj"], withSalary.loc[~did, "sal_adj"])

Internships are both the most common (72% of alumni) and the biggest payoff (+$10.4K). Certifications, co-ops and research add about $4.5K to $5K. Hackathons, campus jobs and competitive teams show no significant gap.

In [19]:stuAlum.assign(role=role).groupby("role")[["sal_adj", "months_to_first_job"]].median()

Raw numbers favor leaders (+$2.5K and 0.5 months to a first job, against −$1.6K and 1.6 months with no club). Once internships, GPA, experience types and credentials are held constant, being a leader adds nothing (see the regression below).

In [20]:orgRank[orgRank["n"] >= 30].nlargest(10, "premium")

Three cybersecurity certification bodies lead, each about $10K above similar grads, followed by consulting and software internship employers.

In [21]:ax.boxplot([adj[grp == k] for k in range(11)])

Grades and courses

In [22]:ax.hexbin(withSalary["final_gpa"], withSalary["sal_adj"], gridsize=25, mincnt=1)

Starting pay climbs steadily with GPA: the median goes from $19K below similar grads at a 1.75–2.0 GPA to $15K above at 3.75–4.0. GPA is the strongest factor in the regression (+$13.2K per GPA point).

In [23]:stuAlum.groupby(stuAlum["dfw"].clip(upper=4))[["time_to_degree_years", "sal_adj"]].agg(["mean", "median"])

Alumni with no D, F or W grades finished in 3.6 years on average; with four or more, 5.2 years. Each one adds about 0.16 years (2 months), holding degree level and entry type fixed. The pay gap (+$3.1K against −$10.3K) runs through GPA: with GPA held constant, D/F/W grades don't predict pay.

In [24]:core["grade"].isin(["D", "F", "W"]).groupby(core["course_id"]).mean().nlargest(10)

CMSC 441 Algorithms (7.4%) and CMSC 341 Data Structures (7.0%) end in a D, F or W most often, then Linear Algebra and Operating Systems. Every CS core course is above the 4.1% rate across all courses, so these are the semesters to plan a lighter load around.

In [25]:smf.ols("sal_adj ~ took + final_gpa + internship_count + credential_count + C(track)", data=d.assign(took=took)).fit()

Among CS electives, only three security courses pass the strict check: Digital Forensics (+$4.7K), Applied Cryptography (+$4.3K) and Offensive Security and Penetration Testing (+$3.6K), compared with CS alumni who have the same GPA, internships, credentials and track. Machine learning and cloud courses lean positive but could be chance.

Statistical checks behind each chart title

One factor at a time: Spearman for ordered factors, Mann-Whitney for two-group gaps (salaries are skewed, so no t-tests).

Claim	Test	ρ	p-value	Verdict
More internships → more likely employed full-time	Spearman	0.076	7.5e-05	Supported
More internships → fewer months to first job	Spearman	-0.572	1.2e-186	Supported
More internships → higher salary premium	Spearman	0.378	9.9e-74	Supported
Return offers pay more than other routes	Mann-Whitney		2.0e-49	Supported
More experience types → higher premium	Spearman	0.294	5.3e-44	Supported
Club leaders earn more than members (raw)	Mann-Whitney		8.2e-03	Supported
Higher GPA → higher premium	Spearman	0.413	3.0e-89	Supported

Which factors still matter together (regression)

OLS regressions for alumni with a first job. Each number is the change for one more unit of that factor, holding the others constant. Club leadership drops out: its raw gap came from leaders also interning more and having higher GPAs.

Factor	Salary premium	p	Significant	Months to first job	p	Significant
Internships	+$5,063	3.4e-38	Yes	−1.10	3.2e-114	Yes
Final GPA	+$13,201	6.9e-77	Yes	−0.32	8.3e-05	Yes
Experience types tried	+$703	1.1e-02	Yes	−0.26	4.4e-15	Yes
Club leader (vs no club)	−$145	8.8e-01	No	+0.12	3.1e-01	No
Club member only (vs no club)	−$164	8.3e-01	No	+0.29	8.5e-04	Yes
Credentials	+$1,551	1.8e-05	Yes	−0.07	9.9e-02	No

Job Trajectory

How grads find their first job, and what happens next

Alumni only. First-job facts come from the alumni file; everything after the first job comes from their employment history. From the team's job-hunting notebook.

Landing the first job

In [26]:pd.crosstab(responded["graduation_year"], responded["first_destination"].map(groups), normalize="index")

Every year, most grads are employed full-time about six months out (69% to 83%). 2020 was the hardest year: 64% full-time and 16% still seeking.

In [27]:np.sort(hired["months_to_first_job"].dropna())

Half of hired alumni start within 0.9 months, three quarters within 3.2 months and 90% within 4.8 months. 18% start right at graduation.

In [28]:hired.groupby("first_job_found_via")["months_to_first_job"].agg(["size", "median"])

Return offers are the most common route (29% of first jobs) and by far the fastest, starting a median 0.2 months after graduation. Every other route takes about two months.

In [29]:hired["first_employer_industry"].value_counts(normalize=True)

46% of first jobs are in the federal ecosystem: federal contracting (22%), federal government (12%), and defense & aerospace (11%).

Pay and moving up

In [30]:firstJobs.groupby("region")[["annual_salary_usd", "adjusted"]].median()

On paper, DC and Northern Virginia pay the most. After cost of living, remote jobs pay the most ($83.1K), and Baltimore ($73.8K) edges out DC ($72.1K).

In [31]:raisePct.groupby(jobs["change_type"]).agg(["median", "size"])

Moving up pays: a new employer at a higher level brings a median 21% raise and an internal promotion 18%. Lateral moves bring about 4%.

In [32]:pd.crosstab(yearsIn, level, normalize="index")

Five years after their first job, 38% of the roles alumni start are Senior or higher, up from almost none in the first two years.

Statistical checks behind each chart title

Chi-square for rates, Mann-Whitney for skewed money and time gaps. The four Baltimore comparisons are Bonferroni-corrected (p × 4). Only supported claims appear in the titles and captions.

Claim	Test	p-value	Verdict
2020 grads were less often employed full-time	Chi-square	9.3e-05	Supported
Return offers start faster than other routes	Mann-Whitney	6.0e-108	Supported
DC paychecks are higher than Baltimore's	Mann-Whitney	2.2e-11	Supported
After cost of living, remote beats Baltimore	Mann-Whitney ×4	5.1e-07	Supported
After cost of living, Baltimore beats Washington, DC Metro	Mann-Whitney ×4	3.9e-02	Supported
After cost of living, Baltimore beats Northern Virginia	Mann-Whitney ×4	5.6e-02	Not significant
After cost of living, Baltimore beats Columbia/Howard County, MD	Mann-Whitney ×4	1.8e-01	Not significant
Step-up moves raise pay more than lateral ones	Mann-Whitney	8.1e-232	Supported

hackUMBC 2026 · UMBC alumni data

What makes a student successful?

“Success” means something different to different people. Here, we use salary as the measure of success and look at how much of the variation in alumni salaries each factor explains.

See the findings Browse the datasets

In [1]:first_job["variance_share"].style.format("{:.1%}")

Share of salary variance explained (η², ANOVA) How much of the variation in alumni salaries each factor explains

Factor	Starting salary	Salary vs peers
Major`major`	1.8%	0.3%
First job family`first_family`	19.7%	20.2%
First employer industry`first_industry`	16.1%	9.5%

2%Salary variance explained by major

20%…by first job family

16%…by first employer industry

9.8%Alumni still seeking work

From the dashboard

Major findings

Salary is our measure of success. Each group below tests one factor against it.

Majors and their effect on success

In [33]:a.plot_salary_by_major_degree(majors["alumcareerinfo"])

Most alumni in the dataset only hold a bachelor's degree (2,032 vs 115 with a master's), so the master's comparisons are built on small samples. The typical salary gap between Computer Science and Information Systems is small: CS alumni earn a median of about 3% more, and the two distributions overlap almost entirely.

In [34]:a.plot_salary_trend(majors["trend"])

IS majors close the initial gap by around the 5-year mark. Fewer alumni remain beyond year 8, so the gap at the end of the chart is a less certain result.

Correlation between major (IS = 1) and starting salary: −0.135. Squared, that is 0.018: only about 2% of the variation in starting salary is explained by major.

Salary summary by major and degree

Major	Degree	Alumni	Avg jobs	Avg salary	Max salary
Computer Science	Bachelor of Science	1,252	2.9	$87,581	$205,500
Computer Science	Master of Science	69	2.9	$90,548	$157,000
Information Systems	Bachelor of Science	780	2.7	$85,850	$220,500
Information Systems	Master of Science	46	2.5	$82,233	$165,500

Internships and their effect on success

In [35]:a.plot_salary_by_internships(intern["df_intern"])

There is a large difference in starting salary between alumni with 0 and 3 internships, but the 2- and 3-internship distributions overlap heavily. Salary gains level off after 2 internships.

In [36]:alum.groupby("nIntern")[["months_to_first_job", "sal_adj"]].median()

Internships help on every count: full-time employment rises from 70% with none to 78% with two or more, time to the first job drops from 4.2 to 0.3 months, and pay goes from $10.0K below similar grads to $8.5K above with three or more.

In [37]:a.plot_return_offers(intern["offer_table"])

At every internship count, alumni whose first job was a return offer earned a higher median starting salary: +$5,250 with 1 internship and +$6,750 with 2 or 3. Return offers also become much more common with more internships (25% with 1, 47% with 2, 66% with 3), so part of the value of more internships is a better chance of converting one into a higher-paying offer.

Outcomes by internship count

Internships	Alumni	Median first-job salary	Median months to hire	Employed full-time	Return offer
0.0	773	$67,750	4.2	58%	0%
1.0	1,368	$78,750	0.9	62%	17%
2.0	837	$83,750	0.3	68%	34%
3.0	212	$87,500	0.3	65%	47%
4.0	10	$70,500	0.2	60%	40%

Return offer premium

Internships	Median, no offer	Median, return offer	Premium	Alumni, no offer	Alumni, offer
1.0	$77,500	$82,750	+$5,250	682	231
2.0	$81,000	$87,750	+$6,750	319	282
3.0	$83,500	$90,250	+$6,750	51	99

First job family and industry

No alumni switched job families or industries, so the first job sets the track. Both factors explain far more salary variation than major does (see the variance table at the top of the page).

In [38]:a.plot_first_job(first_job["career"], first_job["family_summary"], "first_family", "Outcomes by first job family")

Cybersecurity and Infrastructure & Cloud earn much more than peers with the same experience, while IT Support & Operations earns much less.

In [39]:

a.plot_first_job(first_job["career"], first_job["industry_summary"], "first_industry", "Outcomes by first employer industry")

Cloud & Infrastructure and Telecommunications earn much more than Federal Government compared to peers. Cloud & Infrastructure comes out ahead in both views.

Summary tables

First job family	Alumni	Median first salary	Now vs peers
Cybersecurity	342	$91,250	+$11,587
Infrastructure & Cloud	234	$84,750	+$6,029
Machine Learning & AI	177	$83,500	+$1,780
IT Business & Product	274	$75,750	+$443
Software Engineering	481	$75,250	−$5,118
Health IT	109	$75,000	−$4,169
Data & Analytics	425	$72,250	−$9,000
IT Support & Operations	105	$63,000	−$22,958

First employer industry	Alumni	Median first salary	Now vs peers
Cloud & Infrastructure	122	$90,500	+$9,026
Cybersecurity Services	65	$85,250	+$544
Software Products	219	$85,250	+$2,756
Financial Services	103	$84,500	+$4,957
E-Commerce & Retail Tech	125	$81,250	+$638
Consulting & Professional Services	120	$80,875	+$2,883
Telecommunications	53	$80,500	+$8,689
Defense & Aerospace	233	$79,500	+$624
Federal Contracting	482	$78,000	−$3,184
Insurance	59	$76,500	−$2,906
Healthcare & Life Sciences	157	$71,750	−$612
Federal Government	264	$68,250	−$12,662

Unemployment rates

In [40]:a.plot_unemployment(unemp["unemp"], unemp["overall_rate"], by_major)

Unlike the other sections, this includes alumni still seeking work. Overall, 9.8% of alumni in the labor force were still seeking work after graduation, and major makes almost no difference (CS 9.5% vs IS 10.2%).

Internships make a much bigger difference: the rate drops steadily from 12.6% with 0, 10.6% with 1, 7.2% with 2, 4.9% with 3 internships.

Salary and hiring speed leveled off after 2 internships, but unemployment keeps dropping with a 3rd, although that group is smaller (n=144).

Student activities

In [41]:a.plot_experience_share(activities["pct_curr"])

Student organizations are by far the most common experience (66%), followed by internships (44%) and campus jobs (42%). Since internships are tied to higher salaries, faster hiring, and lower unemployment, over half of current students not having one yet is a clear opportunity, though some of this is expected since about half are freshmen or sophomores.

In [42]:a.plot_top_orgs(activities["top10"], activities["baseline"])

3 of the top 4 are certification providers, and most of the rest are internship employers in industries that also paid well as a first job, like Cloud & Infrastructure, Telecommunications, and Cybersecurity Services. This shows correlation, not causation: students who pursue these may already be heading toward higher-paying fields.

Experience types

Internships, co-ops, and hackathons are the most time-intensive (28–40 hours/week), while student organizations, tutoring, and peer mentoring take under 8. Internships and co-ops are mostly paid; student organizations, hackathons, and competitive teams are unpaid. Cells marked n/a mean hours and pay don't apply (certifications).

Experience type	Records	People	Organizations	Avg terms	Avg hours/week	Paid
Student Organization	4,176	3,208	1	3.1	4.0	0%
Internship	4,106	3,012	58	1.0	28.4	87%
Hackathon	3,259	1,489	8	1.0	36.3	0%
Certification	2,845	1,712	17	1.0	n/a	n/a
Campus Job	2,080	2,080	7	3.3	13.0	100%
Undergraduate Research	954	954	7	2.4	10.2	64%
Tutoring	715	715	4	2.5	7.1	100%
Competitive Team	700	700	1	2.0	8.7	0%
Co-op	678	652	58	2.0	40.0	89%
Peer Mentor	546	546	4	2.0	5.3	55%

Method

How we work

Every finding goes through the same five steps. The analysis code lives in analysis.py and the dashboard in app.py.

01
Load
Six CSV files from the hackUMBC 2026 dataset, joined on each person's campus ID.
pandas · read_csv
02
Clean
Turn "Not Applicable" into missing values, convert salaries to numbers, count repeat activities once per person.
pandas
03
Explore
Box plots, hiring curves, and salary trends, one notebook cell per question.
Jupyter · matplotlib
04
Measure
Variance explained by each factor, and salary compared with peers who have the same years of experience.
η² (ANOVA) · numpy
05
Share
Package every chart as a function and serve them in an interactive dashboard.
Streamlit

Data

The datasets

Six CSV files from the hackUMBC 2026 dataset, joined on campus ID.

Dataset	One row is	Rows	Used for
Alumni`alumni.csv`	A graduate, highest degree only	3,200	Majors, internships, unemployment
Employment history`employment_history.csv`	One job spell	6,028	Career salary trend, first job
Student experience`student_experience.csv`	One activity	20,059	Activities, top organizations
Current students`students_current.csv`	A student enrolled in Fall 2026	1,800	Activity participation
Transcripts`transcripts.csv`	One course attempt	140,458	Not used yet
Course catalog`course_catalog.csv`	One course	72	Not used yet
Total		171,617

Day one on campus

What students do outside class

Activities

Organizations

Where it leads

Which in-school choices pay off in job hunting

Internships and job routes

What to do on campus

Grades and courses

How grads find their first job, and what happens next

Landing the first job

Pay and moving up

What makes a student successful?

Major findings

Majors and their effect on success

Internships and their effect on success

First job family and industry

Unemployment rates

Student activities

How we work

Load

Clean

Explore

Measure

Share

The datasets