CRAIN, J.
|PThe defendant appeals a judgment awarding damages in a suit arising out of a plumbing leak that developed inside the plaintiffs residence after the defendant’s employees replaced a water meter at the property. We affirm.
FACTS AND PROCEDURAL HISTORY
On February 8, 2011, at about 2:30 p.m., Ronnie Blake came home from work and found employees of the City of Port Allen working near his driveway. The City was replacing the water meter at Blake’s residence as part of an ongoing project to replace outdated water meters. Blake had lived in the house for over twenty years and had never had any problems with the plumbing. Blake went inside his house, where he observed nothing unusual, and left about ten minutes later to return to work. Approximately two hours later at 4:30 p.m., Blake arrived home from work and saw water running down his driveway. The water was flowing through the doorways and underneath the house where it rests on the slab. Blake attempted to open the side door but encountered some resistance, which he soon discovered was standing water and sheetrock that had fallen from the ceiling. Once inside, he saw water coming through the ceiling of his kitchen from a joint in a pipe in the attic. Blake immediately turned off the water.
[785]*785Blake’s entire residence had flooded, including the kitchen, both bedrooms, and the living room. Photographs admitted into evidence at the trial show extensive water damage to the house and furnishings. The water also destroyed family photographs of Blake’s children. He reported the incident to the City and hired a plumber to replace the pipe the next day. Blake later filed suit against the City, alleging that the City failed to follow proper shut-down procedures and thereby caused the failure of the joint in the pipe.1
|sThe claim proceeded to a bench trial, where Blake presented his testimony along with the testimony of Dilton K. Anderson, a licensed civil engineer, who was accepted by the trial court as an expert in that field. Anderson, who has experience in commercial and residential design of plumbing systems, inspected the section of the pipe that leaked, visited Blake’s residence, and reviewed the pleadings and discovery. At the conclusion of his investigation, Anderson determined that the pipe union had failed because the City employees did not properly turn the water off and on when replacing the water meter. According to Anderson, the employees turned off the water too quickly, creating a vacuum in the line in Blake’s residence, which caused the rubber fitting in the union to collapse; and, after replacing the meter, turned the water on too quickly, which caused a rapid increase in the water pressure.
The City countered with the testimony of Aaron P. Landry, the City’s supervisor of water and gas, and Jeremy C. Hoff-pauir, a mechanical engineer. Landry confirmed the procedure that City employees are supposed to follow when replacing water meters. After checking to see if anyone is home, the employees turn off the water by closing the valve located at the meter. The employees then replace the meter, turn on an outside faucet at the house, and slowly turn the water back on by opening the valve at the meter. They ensure the meter is working and let the air “bleed” out of the system before turning off the outside faucet. Significantly, the final step in the protocol is to “[m]ake sure nothing is leaking” by looking at the meter to see if it indicates any movement of water through the line. The meter has two gauges useful for that purpose: the regular gauge, which measures water in gallons, and a more sensitive “cheater gauge,” which contains a dial that will spin if any water, even a drip, passes through the line.
Landry testified that the workers finished the work at Blake’s residence at about 3:00 p.m. Although he was not present when the work was performed, 14Landry stated that he had no reason to believe the employees deviated from the City’s protocol when they replaced Blake’s water meter. The City employees who performed the work were not called as witnesses, but the parties stipulated that, if called, they would testify that they had followed the protocol when they replaced the water meter.
Landry also described his visit to Blake’s residence the morning after the leak, when he saw the water damage and confirmed that “[everything was still wet.” The pipe had not yet been replaced, and Landry observed the joint where the leak occurred. With Blake’s assistance, Landry attempted to reproduce the leak by turning the water on, but the joint did not leak. Nevertheless, based upon the condition of the area around the joint and the damage to the ceiling beneath it, Landry agreed that it was “pretty obvious” that the fitting had leaked at some point recently.
[786]*786Under cross-examination, Landry was further questioned about the City’s protocol for detecting leaks after replacing a meter, as follows:
Q Now, part of the procedure is that the employees are suppose[d] to look at the cheater dial on the water meter to be sure that there’s no leak?
A Yes, sir.
Q And they’re suppose[d] to observe that cheater dial long enough so that if there’s a leak it should show up because the meter’s going to be running, right? A Yes, sir.
Q And it’s gonna show that there’s a flow of water into the house and you know that there’s a problem?
A Yes, sir.
Q Well, in this particular case if they had looked at the cheater valve or the cheater dial they should’ve seen a leak the size of the one that collapsed the ceiling in this house, shouldn’t they?
* * *
|SA They had told me they didn’t observe any leak.
THE COURT: If there had been a leak of that nature would the cheater valve had shown that?
A The big meter would’ve shown that one.
THE COURT: Okay. Both of them would have?
A Both of them would’ve shown it, yes, sir.
[[Image here]]
THE COURT: And if they see something then they’re suppose[d] to turn it back off and figure out why?
A Yes, sir.
[[Image here]]
Q But here we have a leak that was big enough that not only would the cheater valve show it, but the regular water meter valve would’ve show[n] it?
A Yes, sir, it should’ve.
Q And it didn’t get seen?
They said they did not observe a leak. And, you know, when they turned everything back on they didn’t observe it.
Q Do you know for a fact that they looked?
A They probably did.
[[Image here]]
Q Okay. So a leak the size of this one would’ve been several tens of gallons?
A Seems like it.
Q So it should’ve been obvious?
A Yes, sir, it should’ve been.
The City’s next witness, Hoffpauir, was accepted by the trial court as an expert in mechanical engineering. After inspecting the damaged coupling and the plumbing in Blake’s attic, and reviewing photographs of the damage as well as Anderson’s report and deposition, Hoffpauir concluded that the leak was not Rcaused by turning the water off and on too quickly, but, instead, was due to the age of the gasket in the coupling and a bend in the pipe near the site of the leak. Hoffpauir surmised that the gasket, due to its condition, shifted when the pressure fluctuated in the line as the water was turned off and on during the replacement of the water meter. He further explained that the gasket could have moved back into its proper position when the pressure fluctuated again as Blake turned off the water after he arrived home and Landry turned on the water the following morning during the attempt to reproduce the leak. According to Hoffpauir, that was the only explanation for the leak being present on the afternoon of February 8, 2011, but not the next morning when Landry turned the water back on. Hoff-pauir agreed that a substantial amount of water had leaked from the coupling, and [787]*787he had no evidence that it leaked at any point other than the day the water meter was replaced.
After taking the matter under advisement, the trial court issued written' reasons finding in favor of Blake and awarding him damages of $20,000.00. The court found, in pertinent part:
The testimony and evidence established that a “compression fitting” in the attic failed when the water was turned back on, resulting in the leak and the resulting damage. Both plaintiffs witnesses and defendant’s witnesses agree that there are dangers resulting in increased and decreased pressures when water is turned off and on at a home. Additionally, the City workers were aware of these dangers. The [C]ity witnesses testified that there was a procedure for turning off and on water lines and that they followed those procedures.
However, they also testified that the water meter has two indicators of water flow. The main meter shows movement of 10 gallon increments. There is also a smaller indicator that does not register an amount but will indicate ANY flow of water through the meter. The testimony of the defendant’s witnesses that there was no water flowing when they left is inconsistent with the physical evidence. The plaintiff left his home at 2:30 p.m. while the [C]ity was still changing the meter. He returned home, at 4:30 p.m. and his home was flooded. It would have been physically impossible for that amount of water to leak and not be noticed by anyone monitoring the water meter.
17Either the workers failed to follow procedures and monitor the meter or the meter was defective. Either scenario results in negligence and fault on the part of the [C]ity for the damage that occurred to the plaintiffs residence.
The trial court signed a judgment in accordance with those reasons on November 20, 2013. The City appealed and— asserts three assignments of error: (1) the trial court erred in finding that Blake proved (a) that the City breached a duty owed to him, and (b) legal causation between the City employees’ actions and the alleged flooding; (2) the trial court erred in not placing comparative fault on Blake for the flooding and ensuing damages; and (3) the trial court erred in finding that Blake proved he incurred $20,000.00 in damages.
DISCUSSION
Louisiana courts have adopted a duty-risk analysis in determining whether liability exists under the facts of a particular case. Under this analysis, a plaintiff must prove five separate elements: (1) the defendant had a duty to conform his or her conduct to a specific standard of care, (2) the defendant failed to conform his or her conduct to the appropriate standard of care, (3) the defendant’s substandard conduct was a cause-in-fact of the plaintiffs injuries, (4) the defendant’s substandard conduct was a legal cause of the plaintiffs injuries, and (5) actual damages. Brewer v. J.B. Hunt Transport, Inc., 09-1408 (La.3/16/10), 35 So.3d 230, 240.
The City contends that the trial court erred in finding that the City breached a duty owed to Blake and in finding legal causation between any such breach and Blake’s damages. Whether a duty is owed is a question of law; whether a defendant has breached a duty is a question of fact. Brewer, 35 So.3d at 240; Mundy v. Department of Health and Human Resources, 620 So.2d 811, 813 (La.1993). Legal or proximate cause is ultimately a question of policy as to whether the particular risk falls within the scope of the duty. See Roberts v. Benoit, 605 So.2d 1032, [788]*7881044 (La.1991). Although the determination of legal cause involves a purely legal question, it depends on factual determinations of foreseeability and ease of association. Rando v. Anco Insulations Inc., 08-1163 (La.5/22/09), 16 So.3d 1065, 1088.
A reviewing court may not set aside a trial court’s finding of fact in the absence of manifest error or unless it is clearly wrong, and where there is conflict in the testimony, inferences of fact should not be disturbed upon review, even though the reviewing court may feel that its own evaluations and inferences are. as reasonable. Hanks v. Entergy Corporation, 06-477 (La.12/18/06), 944 So.2d 564, 580; Stobart v. State, Through DOTD, 617 So.2d 880, 882 (La.1993). In order to reverse a trial court’s determination of a fact, a reviewing court must review the record in its entirety and (1) find a reasonable factual basis does not exist for the finding, and (2) further determine the record establishes the factfinder is clearly wrong or manifestly erroneous. Hanks, 944 So.2d at 580; Stobart, 617 So.2d at 882. The issue to be resolved by the reviewing court is not whether the factfinder was right or wrong, but whether the factfinder’s conclusion was a reasonable one. Hanks, 944 So.2d at 580; Stobart, 617 So.2d at 882.
In support of this assignment of error, the City primarily argues that its expert’s explanation of the cause of the leak is more credible than the explanation provided by Blake’s expert. The trial court’s written reasons indicate that the court did not find it necessary to resolve the conflict in the experts’ testimony about the cause of the leak. Instead, the court relied upon evidence establishing that the leak developed while the City employees were replacing the water meter; and the leak was not detected in accordance with the City’s protocol, either because the workers did not check the water meter before leaving, or, if the meter was checked, it was not functioning properly. After a thorough review of the record, we find the evidence reasonably supports that conclusion.
|aIn an action to recover damages for injuries allegedly caused by another’s negligence, the plaintiff has the burden of proving negligence on the part of the defendant by a preponderance of the evidence. Hanks, 944 So.2d at 578. Proof is sufficient to constitute a preponderance when the entirety of the evidence, both direct and circumstantial, shows the fact sought to be proved is more probable than not. Hanks, 944 So.2d at 578; Cangelosi v. Our Lady of the Lake Regional Medical Center, 564 So.2d 654, 664 (La.1989). A fact established by direct evidence is one which has been testified to by witnesses as having come under the cognizance of their senses. Hanks, 944 So.2d at 578; Cangelosi, 564 So.2d at 664. Circumstantial evidence, on the other hand, is evidence of one fact, or of a set of facts, from which the existence of the fact to be determined may reasonably be inferred. Hanks, 944 So.2d at 578-579; Cangelosi, 564 So.2d at 664-65.
The trial testimony indicates that the leak began after the water meter was replaced by the City’s employees. Blake testified that the house appeared fine when he was there at 2:30 p.m., and the City employees were still working. When he returned approximately two hours later, the City employees were gone, and his house was inundated with water leaking from the pipe in, the attic above his kitchen. Landry confirmed that he visited Blake’s residence the next morning, and it was obvious that the fitting had recently leaked. The ceiling underneath the fitting had collapsed, and everything was still wet. Both expert witnesses agreed that the leak began when the water meter was replaced.
[789]*789Landry’s testimony also established that the City’s protocol required the workers to ensure that no leaks were present when the work was completed and the water was turned back on. This was accomplished by looking at the gauges on the replacement meter to see if any water was moving through the line. If a leak was detected, the employees were supposed to turn off the water and determine the source of the problem. According to Landry, a leak of the size that developed at |inBlake’s residence would have been detected by both gauges on the meter and should have been obvious. Although the parties stipulated that the City employees would testify that they followed the proper procedure, the plaintiff did not stipulate that the employees did, in fact, follow the procedure. The evidence produced at trial, both direct and circumstantial, provides a reasonable basis for the trial court’s conclusion that the City employees did not follow the proper procedure to check for a leak before leaving the property, or, alternatively, that the replacement meter was not • functioning properly and failed to reveal the leak.
In this same assignment of error, the City references the trial court’s finding of legal causation; however, the City’s argument — that the pipe union could have failed regardless of the actions by its employees — is directed at the trial court’s finding of cause-in-fact. A trial court’s determination that a defendant’s conduct was a cause-in-fact of the plaintiffs damages is a factual determination subject to the manifest error standard of review. Toston v. Pardon, 03-1747 (La.4/23/04), 874 So.2d 791, 799-800. The determination to be made is whether the harm would have occurred but for the defendant’s alleged substandard conduct, or, when concurrent causes are involved, whether the defendant’s conduct was a substantial factor in bringing about the harm. Granger v. Christus Health Central Louisiana, 12-1892 (La.6/28/13), 144 So.3d 736, 766; Roberts v. Rudzis, 13-0538 (La.App. 1 Cir. 5/28/14), 146 So.3d 602, writ denied, 14-1369 (La.10/3/14), 149 So.3d 797.
The City argues that the leak could have developed even in the absence of any negligent conduct by its employees; however, the trial court’s judgment was based on the City’s failure to detect the leak and turn off the water. The extensive water damage to Blake’s residence would not have occurred but for the failure to detect the leak and immediately turn off the water. Accordingly, the trial court was fynot manifestly erroneous in finding that the substandard conduct was a cause-in-fact of Blake’s damages. The City’s first assignment of error has no merit.2
In its second assignment of error, the City asserts that the trial court erred in not placing comparative fault on Blake for the water damage, because “Blake was responsible for the installation, maintenance and repair of his private water service line, and ... its substandard and deteriorated condition was a contributory, [if] not [the] sole, cause of the water leak and flooding damages.”
The determination and allocation of comparative fault are factual inquiries that will not be disturbed absent manifest error or unless the trial court was clearly wrong. Trinh ex rel. Tran v. Dufrene Boats, Inc., 08-0824 (La.App. 1 Cir. 1/22/09), 6 So.3d 830, 844, writs denied, 09-0406, 09-0411 (La.4/13/09), 5 So.3d 166, [790]*790cert. denied, 558 U.S. 875, 130 S.Ct. 228, 175 L.Ed.2d 128 (2009). The trier of fact is owed great deference in its allocation of fault, and the trial court’s judgment should be affirmed unless it is manifestly erroneous or clearly wrong. See Trinh, 6 So.3d at 844. The defendant bears the burden of proving comparative fault by a preponderance of the evidence. See Granger v. United Home Health Care, 13-0910 (La.App. 1 Cir. 6/19/14), 145 So.3d 1071, 1078; Trinh, 6 So.3d at 844.
Prior to this incident, Blake had lived in the house for over twenty years and had never experienced any problems with the plumbing. When he discovered the leak, Blake immediately turned off the water. The expert witnesses disagreed about the condition of the gasket in the joint and what role, if any, the bend in the pipe played in causing the leak. Anderson testified that he did not see any deficiencies in the pipe. The gasket in the joint, although old, maintained its 112integrity and appeared “functionable,” and he did not think the bend in the pipe contributed to the failure of the joint. In contrast, Hoff-pauir believed the age of the gasket and the bend in the pipe both contributed to the failure of the joint.
Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be manifestly erroneous. Adams v. Rhodia, Inc., 07-2110 (La.5/21/08), 983 So.2d 798, 806. Further, where the findings are based on determinations regarding the credibility of witnesses, the manifest error standard demands great deference to the findings of fact. Adams, 983 So.2d at 806-807. Indeed, where the factfinder’s determination is based on its decision to credit the testimony of one of two or more witnesses, that finding can virtually never be manifestly erroneous. Adams, 983 So.2d at 807. This rule applies equally to the evaluation of expert testimony, including the evaluation and resolution of conflicts in expert testimony. Adams, 983 So.2d at 807. The evidence presented to the trial court supported two permissible views of the condition of the plumbing and what role, if any, that condition played in the leak. The trial court’s choice between those two views cannot be manifestly erroneous. Furthermore, the City did not produce any evidence that Blake knew or should have known of any alleged defects in the home’s plumbing. The trial court’s determination that no comparative fault should be allocated to Blake is reasonably supported by the record and was not manifestly erroneous.
In its final assignment of error, the City contends that the trial court erred in finding that Blake proved $20,000.00 in damages. Arguing that the evidence of special damages amounted to only $2,389.12, the City surmises that the balance of the trial court’s award must be general damages, which the City contends are not appropriate in this case. We need not address whether general damages are appropriate in this case, because the record contains sufficient evidence to support an award of $20,000.00 in property damages caused'by the leak.
|UiIn a suit for damages, it is the plaintiffs burden to prove the damage he suffered as a result of the defendant’s fault. Wainwright v. Fontenot, 00-0492 (La.10/17/00), 774 So.2d 70, 77. Generally, when a person sustains property damage due to the fault of another, he is entitled to recover the cost of restoration that has been or may be reasonably incurred, or he may elect to recover damages based upon the difference between the value of the property before and after the harm. See Roman Catholic Church of Archdiocese of New Orleans v. Louisiana Gas Service Company, 618 So.2d 874, 879 (La.1993); Kahl v. Luster, 11-2332 (La.App. 1 Cir. [791]*79112/28/12), 110 So.3d 1101, 1106. On appellate review, damage awards will be disturbed only when there has been a clear abuse of the trier of fact’s discretion. Hirstius v. BellSouth Telecommunications, Inc., 12-2104 (La.App. 1 Cir. 8/14/13), 123 So.3d 276, 280, writ denied, 13-2709 (La.2/7/14), 131 So.3d 868.
The photographs admitted into evidence show the extensive water damage to Blake’s house. Blake testified that the water damaged multiple rooms in the residence, and he specifically addressed each expense he incurred to replace the damaged items. The work and related expenses included the removal and replacement of drywall, flooring, ceramic tile, and cabinets, as well as expenses for painting, electrical work, storage, and hotel accommodations during the repair and restoration of the house. The total expenses confirmed by Blake’s testimony total well in excess of the trial court’s award of $20,000.00.
The City points out that the invoices and receipts admitted into evidence reflect only $2,389.12 of the expenses Blake incurred in repairing the damage to his home. Blake attempted to introduce invoices or canceled checks for most of the remaining expenses, but the trial court, acting on an objection by the City, did not allow the documents into evidence, because Blake’s counsel did not provide the information to the City’s counsel in response to discovery prior to trial. The |14City relies on the exclusion of that evidence to argue, that Blake’s recovery should be limited to $2,389.12, the amount reflected in the invoices and receipts that were admitted into evidence. However, while the trial court maintained the City’s objection to the introduction of documents that had not been exchanged during discovery, the City did not object to Blake’s testimony confirming the amount he was charged and paid for each item of the repair work.
A trial court does not abuse its discretion in awarding property damages based upon a party’s uncontradicted testimony, admitted without objection, establishing the extent of the damage. See Coleman v. Victor, 326 So.2d 344, 349 (La.1976) (affirming the court award of $800 in repair expenses based upon the plaintiff’s uncontradicted testimony); Malbrough v. Davidson, 219 So.2d 313, 316 (La.App. 1 Cir.1969) (affirming the trial court’s property damage awards based upon a party’s testimony, without objection, establishing the loss of the property and the cost to acquire one item (a motorbike) and the cost to replace other items (clothes)); Smith v. English, 586 So.2d 583, 592 (La.App. 2 Cir.), writ denied, 590 So.2d 80 (La.1991) (affirming the trial court’s property damage award for a boat based upon the plaintiff’s testimony identifying the amount of money he invested in the boat); Tubre v. State, Department of Transportation and Development, 96-1194 (La.App. 3 Cir. 4/2/97), 693 So.2d 1190, 1194, writs denied, 97-1913, 97-1916 (La.11/21/97), 703 So.2d 1307 (affirming the trial court’s property damage award based upon the plaintiffs uncontested testimony establishing the purchase price of the property (a car) paid six months prior to the accident).
In Coleman, the plaintiffs vehicle was damaged in an accident, and the trial court awarded $800.00 for repair costs. Coleman, 326 So.2d at 346. On appeal, this court reversed the property damage award because the “[o]nly evidence regarding the amount necessary to repair the [vehicle] was the self-serving, hearsay 11 .¡testimony of said owner” stating the amount of a repair estimate that was not admitted into evidence. Coleman v. Victor, 314 So.2d 412, 415 (La.App. 1 Cir.1975). The supreme court reversed and reinstated the [792]*792trial court’s award, finding that the plaintiffs testimony, “whether self-serving or hearsay ... became competent evidence” when it was elicited by the defendant during cross-examination, and was “relevant and material and uncontradicted.” Coleman, 326 So.2d at 349. In addition to finding that the defendant effectively waived any right to complain about the competency of the evidence, the supreme court recognized:
Moreover, uncontradicted hearsay testimony of a single witness, admitted without objection, may properly be considered and given probative value.... Likewise, self-serving declarations, admitted without objection, may be considered.
Therefore, Lillie Coleman’s testimony regarding the cost of repair was properly considered and given probative value by the trial judge. We cannot say that he erred in concluding that this uncon-tradicted evidence was sufficient to establish plaintiff-owner’s claim for her property damage. Accordingly, we hold that the court of appeal erred in denying the $800 award for property damage to Lillie Coleman.
Coleman, 326 So.2d at 349 (citation omitted).
We recognize that a plaintiff bears the burden of proving with legal certainty every item of damages and that the plaintiffs own uncorroborated “personal estimate” of loss is insufficient to satisfy that burden. See Tudor Chateau Creole Apartments Partnership v. D.A. Exterminating Co., Inc., 96-0951 (La.App. 1 Cir. 2/14/97), 691 So.2d 1259, 1264; Anderson v. Heck, 554 So.2d 695, 704-05 (La.App. 1 Cir.1989), writ denied, 558 So.2d 605 (La.1990), cert. denied, 498 U.S. 846, 111 S.Ct. 132, 112 L.Ed.2d 100 (1990). However, in the present case, as in Coleman, Malbrough, Smith, and Tubre, the trial court’s damage award was not based upon the plaintiffs personal estimate, nor is the extent of the damage uncorroborated.
|1(iBIake’s testimony established the specific amounts he paid, respectively, to replace the drywall, flooring, cabinets, and ceramic tile damaged by the water from the leak, as well as related expenses for electrical work, painting, and storage. The amounts were not conjecture based upon his “personal estimate” of the loss. To the contrary, they reflect the specific sums that Blake was actually charged and paid for each of those elements of the work. The damage necessitating that work was corroborated by numerous photographs, and the City did not offer any evidence contradicting the amount of the expenses or the need for the repairs. The City likewise did not present any evidence indicating that the costs were unreasonable or excessive. These facts distinguish the present case from Tudor Chateau Creole Apartments Partnership, wherein a party attempted to recover “out-of-pocket expenses” but presented no invoices, canceled checks, or work orders showing that the work was done; and Anderson, wherein a party to a commercial transaction provided only approximations and general testimony concerning the alleged loss resulting from the return of equipment. See Tudor Chateau Creole Apartments Partnership, 691 So.2d at 1264; Anderson, 554 So.2d at 704-705.3
[793]*793The $20,000.00 award by the trial court was significantly less than the total expenses attested to by Blake. As the trial court recognized, “The law allows the court to award minimum reasonable damages for what was proved, even if actual receipts are not available.” Under these circumstances, the trial court did not abuse its discretion or commit manifest error by awarding Blake the sum of $20,000.00.
^CONCLUSION
We affirm the judgment of the trial court signed on November 20, 2013, and assess all costs of appeal in the amount of $1,668.00 to the City of Port Allen.
AFFIRMED.